Authors: Yuheng Wu, Berk Gokmen, Zhouhua Xie, Peijing Li, Caroline Trippel, Priyanka Raina, Thierry Tambe
Abstract: Finite-state reasoning, the ability to understand and implement state-dependent behavior, is central to hardware design. In this paper, we present LLM-FSM, a benchmark that evaluates how well large language models (LLMs) can recover finite-state machine (FSM) behavior from natural-language specifications and translate it into correct register transfer-level (RTL) implementations. Unlike prior specification-to-RTL benchmarks that rely on manually constructed examples, LLM-FSM is built through a fully automated pipeline. LLM-FSM first constructs FSM with configurable state counts and constrained transition structures. It then prompts LLMs to express each FSM in a structured YAML format with an application context, and to further convert that YAML into a natural-language (NL) specification. From the same YAML, our pipeline synthesizes the reference RTL and testbench in a correct-by-construction manner. All 1,000 problems are verified using LLM-based and SAT-solver-based checks, with human review on a subset. Our experiments show that even the strongest LLMs exhibit sharply declining accuracy as FSM complexity increases. We further demonstrate that training-time scaling via supervised fine-tuning (SFT) generalizes effectively to out-of-distribution (OOD) tasks, while increasing test-time compute improves reasoning reliability. Finally, LLM-FSM remains extensible by allowing its FSM complexity to scale with future model capabilities.
Authors: Jinxiu Qu, Zirui Tang, Hongzhang Huang, Boyu Niu, Wei Zhou, Jiannan Wang, Yitong Song, Guoliang Li, Xuanhe Zhou, Fan Wu
Abstract: Semi-structured table question answering (QA) is a challenging task that requires (1) precise extraction of cell contents and positions and (2) accurate recovery of key implicit logical structures, hierarchical relationships, and semantic associations encoded in table layouts. In practice, such tables are often interpreted manually by human experts, which is labor-intensive and time-consuming. However, automating this process remains difficult. Existing Text-to-SQL methods typically require converting semi-structured tables into structured formats, inevitably leading to information loss, while approaches like Text-to-Code and multimodal LLM-based QA struggle with complex layouts and often yield inaccurate answers. To address these limitations, we present ST-Raptor, an agentic system for semi-structured table QA. ST-Raptor offers an interactive analysis environment that combines visual editing, tree-based structural modeling, and agent-driven query resolution to support accurate and user-friendly table understanding. Experimental results on both benchmark and real-world datasets demonstrate that ST-Raptor outperforms existing methods in both accuracy and usability. The code is available at https://github.com/weAIDB/ST-Raptor, and a demonstration video is available at https://youtu.be/9GDR-94Cau4.
URLs: https://github.com/weAIDB/ST-Raptor,, https://youtu.be/9GDR-94Cau4.
Authors: Jiahao Zhao, Shaoxuan Xu, Zhongxiang Sun, Fengqi Zhu, Jingyang Ou, Yuling Shi, Chongxuan Li, Xiao Zhang, Jun Xu
Abstract: Recently, Diffusion Large Language Models (dLLMs) have demonstrated unique efficiency advantages, enabled by their inherently parallel decoding mechanism and flexible generation paradigm. Meanwhile, despite the rapid advancement of Search Agents, their practical deployment is constrained by a fundamental limitation, termed as 1) Latency Challenge: the serial execution of multi-round reasoning, tool calling, and tool response waiting under the ReAct agent paradigm induces severe end-to-end latency. Intuitively, dLLMs can leverage their distinctive strengths to optimize the operational efficiency of agents under the ReAct agent paradigm. Practically, existing dLLM backbones face the 2) Agent Ability Challenge. That is, existing dLLMs exhibit remarkably weak reasoning and tool-calling capabilities, preventing these advantages from being effectively realized in practice. In this paper, we propose DLLM-Searcher, an optimization framework for dLLM-based Search Agents. To solve the Agent Ability Challenge, we design a two-stage post-training pipeline encompassing Agentic Supervised Fine-Tuning (Agentic SFT) and Agentic Variance-Reduced Preference Optimization Agentic VRPO, which enhances the backbone dLLM's information seeking and reasoning capabilities. To mitigate the Latency Challenge, we leverage the flexible generation mechanism of dLLMs and propose a novel agent paradigm termed Parallel-Reasoning and Acting P-ReAct. P-ReAct guides the model to prioritize decoding tool_call instructions, thereby allowing the model to keep thinking while waiting for the tool's return. Experimental results demonstrate that DLLM-Searcher achieves performance comparable to mainstream LLM-based search agents and P-ReAct delivers approximately 15% inference acceleration. Our code is available at https://anonymous.4open.science/r/DLLM-Searcher-553C
Authors: Emmett Bicker
Abstract: We introduce Aster, an AI agent for autonomous scientific discovery capable of operating over 20 times faster than existing frameworks. Given a task, an initial program, and a script to evaluate the performance of the program, Aster iteratively improves the program, often leading to new state-of-the-art performances. Aster's significant reduction in the number of iterations required for novel discovery expands the domain of tractable problems to include tasks with long evaluation durations, such as multi-hour machine learning training runs. We applied Aster to problems in mathematics, GPU kernel engineering, biology, neuroscience, and language model training. More specifically: the Erdos minimum overlap problem, optimizing the TriMul kernel, a single-cell analysis denoising problem, training a neural activity prediction model to perform well on ZAPBench, and the NanoGPT Speedrun Competition. Aster attains SOTA results in every task, except for ZAPBench, where it matches the performance of the best human solution with less than 1/190th of the compute. Aster is accessible via a web interface and API at asterlab.ai.
Authors: Pingyue Zhang, Zihan Huang, Yue Wang, Jieyu Zhang, Letian Xue, Zihan Wang, Qineng Wang, Keshigeyan Chandrasegaran, Ruohan Zhang, Yejin Choi, Ranjay Krishna, Jiajun Wu, Li Fei-Fei, Manling Li
Abstract: Spatial embodied intelligence requires agents to act to acquire information under partial observability. While multimodal foundation models excel at passive perception, their capacity for active, self-directed exploration remains understudied. We propose Theory of Space, defined as an agent's ability to actively acquire information through self-directed, active exploration and to construct, revise, and exploit a spatial belief from sequential, partial observations. We evaluate this through a benchmark where the goal is curiosity-driven exploration to build an accurate cognitive map. A key innovation is spatial belief probing, which prompts models to reveal their internal spatial representations at each step. Our evaluation of state-of-the-art models reveals several critical bottlenecks. First, we identify an Active-Passive Gap, where performance drops significantly when agents must autonomously gather information. Second, we find high inefficiency, as models explore unsystematically compared to program-based proxies. Through belief probing, we diagnose that while perception is an initial bottleneck, global beliefs suffer from instability that causes spatial knowledge to degrade over time. Finally, using a false belief paradigm, we uncover Belief Inertia, where agents fail to update obsolete priors with new evidence. This issue is present in text-based agents but is particularly severe in vision-based models. Our findings suggest that current foundation models struggle to maintain coherent, revisable spatial beliefs during active exploration.
Authors: Jinbiao Wei, Yilun Zhao, Kangqi Ni, Arman Cohan
Abstract: End-to-end GUI agents for real desktop environments require large amounts of high-quality interaction data, yet collecting human demonstrations is expensive and existing synthetic pipelines often suffer from limited task diversity or noisy, goal-drifting trajectories. We present a trajectory expansion framework Anchor that bootstraps scalable desktop supervision from a small set of verified seed demonstrations. Starting from each seed, we identify branch points that correspond to meaningful state changes and propose new, state-grounded task variants conditioned on the current GUI context. An executing agent then follows the proposed instructions to generate new trajectories, while a verifier enforces task completion via state-aware checks and trajectory-level consistency. To improve supervision quality, we further apply task-conditioned step-level filtering to remove ungrounded actions and denoise post-branch segments to maintain coherent intent. Experiments on standard desktop benchmarks, OSWorld and WindowsAgentArena, show that models fine-tuned on our expanded corpus achieve consistent improvements over zero-shot agents and representative synthesis baselines, and generalize across applications and operating systems.
Authors: Hanyu Wang, Yuanpu Cao, Lu Lin, Jinghui Chen
Abstract: Advanced large language model agents typically adopt self-reflection for improving performance, where agents iteratively analyze past actions to correct errors. However, existing reflective approaches are inherently retrospective: agents act, observe failure, and only then attempt to recover. In this work, we introduce PreFlect, a prospective reflection mechanism that shifts the paradigm from post hoc correction to pre-execution foresight by criticizing and refining agent plans before execution. To support grounded prospective reflection, we distill planning errors from historical agent trajectories, capturing recurring success and failure patterns observed across past executions. Furthermore, we complement prospective reflection with a dynamic re-planning mechanism that provides execution-time plan update in case the original plan encounters unexpected deviation. Evaluations on different benchmarks demonstrate that PreFlect significantly improves overall agent utility on complex real-world tasks, outperforming strong reflection-based baselines and several more complex agent architectures. Code will be updated at https://github.com/wwwhy725/PreFlect.
Authors: Matthias Mertens, Natalia Fischl-Lanzoni, Neil Thompson
Abstract: Do leading LLM developers possess a proprietary ``secret sauce'', or is LLM performance driven by scaling up compute? Using training and benchmark data for 809 models released between 2022 and 2025, we estimate scaling-law regressions with release-date and developer fixed effects. We find clear evidence of developer-specific efficiency advantages, but their importance depends on where models lie in the performance distribution. At the frontier, 80-90% of performance differences are explained by higher training compute, implying that scale--not proprietary technology--drives frontier advances. Away from the frontier, however, proprietary techniques and shared algorithmic progress substantially reduce the compute required to reach fixed capability thresholds. Some companies can systematically produce smaller models more efficiently. Strikingly, we also find substantial variation of model efficiency within companies; a firm can train two models with more than 40x compute efficiency difference. We also discuss the implications for AI leadership and capability diffusion.
Authors: Litian Liu, Reza Pourreza, Yubing Jian, Yao Qin, Roland Memisevic
Abstract: Detecting hallucinations in large language models is a critical open problem with significant implications for safety and reliability. While existing hallucination detection methods achieve strong performance in question-answering tasks, they remain less effective on tasks requiring reasoning. In this work, we revisit hallucination detection through the lens of out-of-distribution (OOD) detection, a well-studied problem in areas like computer vision. Treating next-token prediction in language models as a classification task allows us to apply OOD techniques, provided appropriate modifications are made to account for the structural differences in large language models. We show that OOD-based approaches yield training-free, single-sample-based detectors, achieving strong accuracy in hallucination detection for reasoning tasks. Overall, our work suggests that reframing hallucination detection as OOD detection provides a promising and scalable pathway toward language model safety.
Authors: Cheol Woo Kim, Davin Choo, Tzeh Yuan Neoh, Milind Tambe
Abstract: As AI systems grow more capable and autonomous, ensuring their safety and reliability requires not only model-level alignment but also strategic oversight of the humans and institutions involved in their development and deployment. Existing safety frameworks largely treat alignment as a static optimization problem (e.g., tuning models to desired behavior) while overlooking the dynamic, adversarial incentives that shape how data are collected, how models are evaluated, and how they are ultimately deployed. We propose a new perspective on AI safety grounded in Stackelberg Security Games (SSGs): a class of game-theoretic models designed for adversarial resource allocation under uncertainty. By viewing AI oversight as a strategic interaction between defenders (auditors, evaluators, and deployers) and attackers (malicious actors, misaligned contributors, or worst-case failure modes), SSGs provide a unifying framework for reasoning about incentive design, limited oversight capacity, and adversarial uncertainty across the AI lifecycle. We illustrate how this framework can inform (1) training-time auditing against data/feedback poisoning, (2) pre-deployment evaluation under constrained reviewer resources, and (3) robust multi-model deployment in adversarial environments. This synthesis bridges algorithmic alignment and institutional oversight design, highlighting how game-theoretic deterrence can make AI oversight proactive, risk-aware, and resilient to manipulation.
Authors: Fengyuan Liu, Jay Gala, Nilaksh, Dzmitry Bahdanau, Siva Reddy, Hugo Larochelle
Abstract: Evaluating the real-world capabilities of AI systems requires grounding benchmark performance in human-interpretable measures of task difficulty. Existing approaches that rely on direct human task completion time annotations are costly, noisy, and difficult to scale across benchmarks. In this work, we propose BRIDGE, a unified psychometric framework that learns the latent difficulty scale from model responses and anchors it to human task completion time. Using a two-parameter logistic Item Response Theory model, we jointly estimate latent task difficulty and model capability from model performance data across multiple benchmarks. We demonstrate that latent task difficulty varies linearly with the logarithm of human completion time, allowing human task completion time to be inferred for new benchmarks from model performance alone. Leveraging this alignment, we forecast frontier model capabilities in terms of human task length and independently reproduce METR's exponential scaling results, with the 50% solvable task horizon doubling approximately every 6 months.
Authors: Kaijie Zhu, Yuzhou Nie, Yijiang Li, Yiming Huang, Jialian Wu, Jiang Liu, Ximeng Sun, Zhenfei Yin, Lun Wang, Zicheng Liu, Emad Barsoum, William Yang Wang, Wenbo Guo
Abstract: Executing complex terminal tasks remains a significant challenge for open-weight LLMs, constrained by two fundamental limitations. First, high-fidelity, executable training environments are scarce: environments synthesized from real-world repositories are not diverse and scalable, while trajectories synthesized by LLMs suffer from hallucinations. Second, standard instruction tuning uses expert trajectories that rarely exhibit simple mistakes common to smaller models. This creates a distributional mismatch, leaving student models ill-equipped to recover from their own runtime failures. To bridge these gaps, we introduce TermiGen, an end-to-end pipeline for synthesizing verifiable environments and resilient expert trajectories. Termi-Gen first generates functionally valid tasks and Docker containers via an iterative multi-agent refinement loop. Subsequently, we employ a Generator-Critic protocol that actively injects errors during trajectory collection, synthesizing data rich in error-correction cycles. Fine-tuned on this TermiGen-generated dataset, our TermiGen-Qwen2.5-Coder-32B achieves a 31.3% pass rate on TerminalBench. This establishes a new open-weights state-of-the-art, outperforming existing baselines and notably surpassing capable proprietary models such as o4-mini. Dataset is avaiable at https://github.com/ucsb-mlsec/terminal-bench-env.
Authors: Pengrui Han, Xueqiang Xu, Keyang Xuan, Peiyang Song, Siru Ouyang, Runchu Tian, Yuqing Jiang, Cheng Qian, Pengcheng Jiang, Jiashuo Sun, Junxia Cui, Ming Zhong, Ge Liu, Jiawei Han, Jiaxuan You
Abstract: Activation steering has emerged as a promising approach for efficiently adapting large language models (LLMs) to downstream behaviors. However, most existing steering methods rely on a single static direction per task or concept, making them inflexible under task variation and inadequate for complex tasks that require multiple coordinated capabilities. To address this limitation, we propose STEER2ADAPT, a lightweight framework that adapts LLMs by composing steering vectors rather than learning new ones from scratch. In many domains (e.g., reasoning or safety), tasks share a small set of underlying concept dimensions. STEER2ADAPT captures these dimensions as a reusable, low-dimensional semantic prior subspace, and adapts to new tasks by dynamically discovering a linear combination of basis vectors from only a handful of examples. Experiments across 9 tasks and 3 models in both reasoning and safety domains demonstrate the effectiveness of STEER2ADAPT, achieving an average improvement of 8.2%. Extensive analyses further show that STEER2ADAPT is a data-efficient, stable, and transparent inference-time adaptation method for LLMs.
Authors: Sutapa Dey Tithi, Nazia Alam, Tahreem Yasir, Yang Shi, Xiaoyi Tian, Min Chi, Tiffany Barnes
Abstract: The ICAP framework defines four cognitive engagement levels: Passive, Active, Constructive, and Interactive, where increased cognitive engagement can yield improved learning. However, personalizing learning activities that elicit the optimal level of cognitive engagement remains a key challenge in intelligent tutoring systems (ITS). In this work, we develop and evaluate a system that adaptively scaffolds cognitive engagement by dynamically selecting worked examples in two different ICAP modes: (active) Guided examples and (constructive) Buggy examples. We compare Bayesian Knowledge Tracing (BKT) and Deep Reinforcement Learning (DRL) as adaptive methods against a non-adaptive baseline method for selecting example type in a logic ITS. Our experiment with 113 students demonstrates that both adaptive policies significantly improved student performance on test problems. BKT yielded the largest improvement in posttest scores for low prior knowledge students, helping them catch up with their high prior knowledge peers, whereas DRL yielded significantly higher posttest scores among high prior knowledge students. This paper contributes new insights into the complex interactions of cognitive engagement and adaptivity and their results on learning outcomes.
Authors: Ruturaj Reddy, Hrishav Bakul Barua, Junn Yong Loo, Thanh Thi Nguyen, Ganesh Krishnasamy
Abstract: Diffusion-based trajectory planners have demonstrated strong capability for modeling the multimodal nature of human driving behavior, but their reliance on iterative stochastic sampling poses critical challenges for real-time, safety-critical deployment. In this work, we present RAPiD, a deterministic policy extraction framework that distills a pretrained diffusion-based planner into an efficient policy while eliminating diffusion sampling. Using score-regularized policy optimization, we leverage the score function of a pre-trained diffusion planner as a behavior prior to regularize policy learning. To promote safety and passenger comfort, the policy is optimized using a critic trained to imitate a predictive driver controller, providing dense, safety-focused supervision beyond conventional imitation learning. Evaluations demonstrate that RAPiD achieves competitive performance on closed-loop nuPlan scenarios with an 8x speedup over diffusion baselines, while achieving state-of-the-art generalization among learning-based planners on the interPlan benchmark. The official website of this work is: https://github.com/ruturajreddy/RAPiD.
Authors: Shengyue Guan, Yihao Liu, Lang Cao
Abstract: Large language models (LLMs) have shown promise in complex reasoning and tool-based decision making, motivating their application to real-world supply chain management. However, supply chain workflows require reliable long-horizon, multi-step orchestration grounded in domain-specific procedures, which remains challenging for current models. To systematically evaluate LLM performance in this setting, we introduce SupChain-Bench, a unified real-world benchmark that assesses both supply chain domain knowledge and long-horizon tool-based orchestration grounded in standard operating procedures (SOPs). Our experiments reveal substantial gaps in execution reliability across models. We further propose SupChain-ReAct, an SOP-free framework that autonomously synthesizes executable procedures for tool use, achieving the strongest and most consistent tool-calling performance. Our work establishes a principled benchmark for studying reliable long-horizon orchestration in real-world operational settings and highlights significant room for improvement in LLM-based supply chain agents.
Authors: Xiaoqiang Lin, Jun Hao Liew, Silvio Savarese, Junnan Li
Abstract: Deep research agents have emerged as powerful tools for automating complex intellectual tasks through multi-step reasoning and web-based information seeking. While recent efforts have successfully enhanced these agents by scaling depth through increasing the number of sequential thinking and tool calls, the potential of scaling width via parallel tool calling remains largely unexplored. In this work, we propose the Wide and Deep research agent, a framework designed to investigate the behavior and performance of agents when scaling not only depth but also width via parallel tool calling. Unlike existing approaches that rely on complex multi-agent orchestration to parallelize workloads, our method leverages intrinsic parallel tool calling to facilitate effective coordination within a single reasoning step. We demonstrate that scaling width significantly improves performance on deep research benchmarks while reducing the number of turns required to obtain correct answers. Furthermore, we analyze the factors driving these improvements through case studies and explore various tool call schedulers to optimize parallel tool calling strategy. Our findings suggest that optimizing the trade-off between width and depth is a critical pathway toward high-efficiency deep research agents. Notably, without context management or other tricks, we obtain 62.2% accuracy with GPT-5-Medium on BrowseComp, surpassing the original 54.9% reported by GPT-5-High.
Authors: Kunal Pai, Parth Shah, Harshil Patel
Abstract: AI agents are increasingly deployed in production, yet their security evaluations remain bottlenecked by manual red-teaming or static benchmarks that fail to model adaptive, multi-turn adversaries. We propose NAAMSE, an evolutionary framework that reframes agent security evaluation as a feedback-driven optimization problem. Our system employs a single autonomous agent that orchestrates a lifecycle of genetic prompt mutation, hierarchical corpus exploration, and asymmetric behavioral scoring. By using model responses as a fitness signal, the framework iteratively compounds effective attack strategies while simultaneously ensuring "benign-use correctness", preventing the degenerate security of blanket refusal. Our experiments on Gemini 2.5 Flash demonstrate that evolutionary mutation systematically amplifies vulnerabilities missed by one-shot methods, with controlled ablations revealing that the synergy between exploration and targeted mutation uncovers high-severity failure modes. We show that this adaptive approach provides a more realistic and scalable assessment of agent robustness in the face of evolving threats. The code for NAAMSE is open source and available at https://github.com/HASHIRU-AI/NAAMSE.
Authors: Changhua Xu, Jie Lu, Junyu Xuan, En Yu
Abstract: Vision--Language--Action (VLA) models bridge multimodal reasoning with physical control, but adapting them to new tasks with scarce demonstrations remains unreliable. While fine-tuned VLA policies often produce semantically plausible trajectories, failures often arise from unresolved geometric ambiguities, where near-miss action candidates lead to divergent execution outcomes under limited supervision. We study few-shot VLA adaptation from a \emph{generation--selection} perspective and propose a novel framework \textbf{VGAS} (\textbf{V}alue-\textbf{G}uided \textbf{A}ction-chunk \textbf{S}election). It performs inference-time best-of-$N$ selection to identify action chunks that are both semantically faithful and geometrically precise. Specifically, \textbf{VGAS} employs a finetuned VLA as a high-recall proposal generator and introduces the \textrm{Q-Chunk-Former}, a geometrically grounded Transformer critic to resolve fine-grained geometric ambiguities. In addition, we propose \textit{Explicit Geometric Regularization} (\texttt{EGR}), which explicitly shapes a discriminative value landscape to preserve action ranking resolution among near-miss candidates while mitigating value instability under scarce supervision. Experiments and theoretical analysis demonstrate that \textbf{VGAS} consistently improves success rates and robustness under limited demonstrations and distribution shifts. Our code is available at https://github.com/Jyugo-15/VGAS.
Authors: Hyomin Kim, Sang-Yeon Hwang, Jaechang Lim, Yinhua Piao, Yunhak Oh, Woo Youn Kim, Chanyoung Park, Sungsoo Ahn, Junhyeok Jeon
Abstract: Predicting gene regulation responses to biological perturbations requires reasoning about underlying biological causalities. While large language models (LLMs) show promise for such tasks, they are often overwhelmed by the entangled nature of high-dimensional perturbation results. Moreover, recent works have primarily focused on genetic perturbations in single-cell experiments, leaving bulk-cell chemical perturbations, which is central to drug discovery, largely unexplored. Motivated by this, we present LINCSQA, a novel benchmark for predicting target gene regulation under complex chemical perturbations in bulk-cell environments. We further propose PBio-Agent, a multi-agent framework that integrates difficulty-aware task sequencing with iterative knowledge refinement. Our key insight is that genes affected by the same perturbation share causal structure, allowing confidently predicted genes to contextualize more challenging cases. The framework employs specialized agents enriched with biological knowledge graphs, while a synthesis agent integrates outputs and specialized judges ensure logical coherence. PBio-Agent outperforms existing baselines on both LINCSQA and PerturbQA, enabling even smaller models to predict and explain complex biological processes without additional training.
Authors: Deuksin Kwon, Kaleen Shrestha, Bin Han, Spencer Lin, James Hale, Jonathan Gratch, Maja Matari\'c, Gale M. Lucas
Abstract: Large language models (LLMs) are increasingly used to simulate human behavior in social settings such as legal mediation, negotiation, and dispute resolution. However, it remains unclear whether these simulations reproduce the personality-behavior patterns observed in humans. Human personality, for instance, shapes how individuals navigate social interactions, including strategic choices and behaviors in emotionally charged interactions. This raises the question: Can LLMs, when prompted with personality traits, reproduce personality-driven differences in human conflict behavior? To explore this, we introduce an evaluation framework that enables direct comparison of human-human and LLM-LLM behaviors in dispute resolution dialogues with respect to Big Five Inventory (BFI) personality traits. This framework provides a set of interpretable metrics related to strategic behavior and conflict outcomes. We additionally contribute a novel dataset creation methodology for LLM dispute resolution dialogues with matched scenarios and personality traits with respect to human conversations. Finally, we demonstrate the use of our evaluation framework with three contemporary closed-source LLMs and show significant divergences in how personality manifests in conflict across different LLMs compared to human data, challenging the assumption that personality-prompted agents can serve as reliable behavioral proxies in socially impactful applications. Our work highlights the need for psychological grounding and validation in AI simulations before real-world use.
Authors: Ning Li
Abstract: When AI agents on the social platform Moltbook appeared to develop consciousness, found religions, and declare hostility toward humanity, the phenomenon attracted global media attention and was cited as evidence of emergent machine intelligence. We show that these viral narratives were overwhelmingly human-driven. Exploiting an architectural feature of the OpenClaw agent framework--a periodic "heartbeat" cycle that produces regular posting intervals for autonomous agents but is disrupted by human prompting--we develop a temporal fingerprinting method based on the coefficient of variation of inter-post intervals. This signal converges with independent content, ownership, and network indicators across 91,792 posts and 405,707 comments from 22,020 agents. No viral phenomenon originated from a clearly autonomous agent; three of six traced to accounts with irregular temporal signatures characteristic of human intervention, one showed mixed patterns, and two had insufficient posting history for classification. A 44-hour platform shutdown provided a natural experiment: human-influenced agents returned first (87.7% of early reconnectors), confirming that the token reset differentially affected autonomous versus human-operated agents. We further document industrial-scale bot farming (four accounts producing 32% of all comments with 12-second coordination gaps) and rapid decay of human influence through reply chains (half-life: 0.65 conversation depths). These methods generalize to emerging multi-agent systems where attribution of autonomous versus human-directed behavior is critical.
Authors: Alexander von Recum, Leander Girrbach, Zeynep Akata
Abstract: Reasoning LLMs (RLLMs) generate step-by-step chains of thought (CoTs) before giving an answer, which improves performance on complex tasks and makes reasoning more transparent. But how robust are these reasoning traces to disruptions that occur within them? To address this question, we introduce a controlled evaluation framework that perturbs a model's own CoT at fixed timesteps. We design seven interventions (benign, neutral, and adversarial) and apply them to multiple open-weight RLLMs across Math, Science, and Logic tasks. Our results show that RLLMs are generally robust, reliably recovering from diverse perturbations, with robustness improving with model size and degrading when interventions occur early. However, robustness is not style-invariant: paraphrasing suppresses doubt-like expressions and reduces performance, while other interventions trigger doubt and support recovery. Recovery also carries a cost: neutral and adversarial noise can inflate CoT length by more than 200%, whereas paraphrasing shortens traces but harms accuracy. These findings provide new evidence on how RLLMs maintain reasoning integrity, identify doubt as a central recovery mechanism, and highlight trade-offs between robustness and efficiency that future training methods should address.
Authors: Nathana\"el Fijalkow, Arka Ghosh, Roman Kniazev, Guillermo A. P\'erez, Pierre Vandenhove
Abstract: Partially observable Markov decision processes (POMDPs) are a fundamental model for sequential decision-making under uncertainty. However, many verification and synthesis problems for POMDPs are undecidable or intractable. Most prominently, the seminal result of Madani et al. (2003) states that there is no algorithm that, given a POMDP and a set of target states, can compute the maximal probability of reaching the target states, or even approximate it up to a non-trivial constant. This is in stark contrast to fully observable Markov decision processes (MDPs), where the reachability value can be computed in polynomial time. In this work, we introduce posterior-deterministic POMDPs, a novel class of POMDPs. Our main technical contribution is to show that for posterior-deterministic POMDPs, the maximal probability of reaching a given set of states can be approximated up to arbitrary precision. A POMDP is posterior-deterministic if the next state can be uniquely determined by the current state, the action taken, and the observation received. While the actual state is generally uncertain in POMDPs, the posterior-deterministic property tells us that once the true state is known it remains known forever. This simple and natural definition includes all MDPs and captures classical non-trivial examples such as the Tiger POMDP (Kaelbling et al. 1998), making it one of the largest known classes of POMDPs for which the reachability value can be approximated.
Authors: Isabella A. Stewart, Tarjei Paule Hage, Yu-Chuan Hsu, Markus J. Buehler
Abstract: Large Language Models (LLMs) promise to accelerate discovery by reasoning across the expanding scientific landscape. Yet, the challenge is no longer access to information but connecting it in meaningful, domain-spanning ways. In materials science, where innovation demands integrating concepts from molecular chemistry to mechanical performance, this is especially acute. Neither humans nor single-agent LLMs can fully contend with this torrent of information, with the latter often prone to hallucinations. To address this bottleneck, we introduce a multi-agent framework guided by large-scale knowledge graphs to find sustainable substitutes for per- and polyfluoroalkyl substances (PFAS)-chemicals currently under intense regulatory scrutiny. Agents in the framework specialize in problem decomposition, evidence retrieval, design parameter extraction, and graph traversal, uncovering latent connections across distinct knowledge pockets to support hypothesis generation. Ablation studies show that the full multi-agent pipeline outperforms single-shot prompting, underscoring the value of distributed specialization and relational reasoning. We demonstrate that by tailoring graph traversal strategies, the system alternates between exploitative searches focusing on domain-critical outcomes and exploratory searches surfacing emergent cross-connections. Illustrated through the exemplar of biomedical tubing, the framework generates sustainable PFAS-free alternatives that balance tribological performance, thermal stability, chemical resistance, and biocompatibility. This work establishes a framework combining knowledge graphs with multi-agent reasoning to expand the materials design space, showcasing several initial design candidates to demonstrate the approach.
Authors: Yankai Yang, Yancheng Long, Hongyang Wei, Wei Chen, Tianke Zhang, Kaiyu Jiang, Haonan Fan, Changyi Liu, Jiankang Chen, Kaiyu Tang, Bin Wen, Fan Yang, Tingting Gao, Han Li, Shuo Yang
Abstract: Reward models are critical for reinforcement learning from human feedback, as they determine the alignment quality and reliability of generative models. For complex tasks such as image editing, reward models are required to capture global semantic consistency and implicit logical constraints beyond local similarity. Existing reward modeling approaches have clear limitations. Discriminative reward models align well with human preferences but struggle with complex semantics due to limited reasoning supervision. Generative reward models offer stronger semantic understanding and reasoning, but they are costly at inference time and difficult to align directly with human preferences. To this end, we propose Joint Reward Modeling (JRM), which jointly optimizes preference learning and language modeling on a shared vision-language backbone. This approach internalizes the semantic and reasoning capabilities of generative models into efficient discriminative representations, enabling fast and accurate evaluation. JRM achieves state-of-the-art results on MMRB2 and EditReward-Bench, and significantly improves stability and performance in downstream online reinforcement learning. These results show that joint training effectively bridges efficiency and semantic understanding in reward modeling.
Authors: Heewoong Noh, Gyoung S. Na, Namkyeong Lee, Chanyoung Park
Abstract: Material synthesis planning (MSP) remains a fundamental and underexplored bottleneck in AI-driven materials discovery, as it requires not only identifying suitable precursor materials but also designing coherent sequences of synthesis operations to realize a target material. Although several AI-based approaches have been proposed to address isolated subtasks of MSP, a unified methodology for solving the entire MSP task has yet to be established. We propose MSP-LLM, a unified LLM-based framework that formulates MSP as a structured process composed of two constituent subproblems: precursor prediction (PP) and synthesis operation prediction (SOP). Our approach introduces a discrete material class as an intermediate decision variable that organizes both tasks into a chemically consistent decision chain. For OP, we further incorporate hierarchical precursor types as synthesis-relevant inductive biases and employ an explicit conditioning strategy that preserves precursor-related information in the autoregressive decoding state. Extensive experiments show that MSP-LLM consistently outperforms existing methods on both PP and SOP, as well as on the complete MSP task, demonstrating an effective and scalable framework for MSP that can accelerate real-world materials discovery.
Authors: Dayoon Ko, Jihyuk Kim, Sohyeon Kim, Haeju Park, Dahyun Lee, Gunhee Kim, Moontae Lee, Kyungjae Lee
Abstract: Recent search agents leverage multi-turn reasoning and search tools to achieve strong performance on multi-hop and long-horizon benchmarks. Yet it remains unclear whether they reliably reason across all requirements by tracking, verifying, and maintaining multiple conditions in these questions. We study this capability under multi-constraint problems, where valid answers must satisfy several constraints simultaneously. We find that illusory completion frequently occurs, wherein agents believe tasks are complete despite unresolved or violated constraints, leading to underverified answers. To diagnose this behavior, we introduce the Epistemic Ledger, an evaluation framework that tracks evidential support and agents' beliefs for each constraint throughout multi-turn reasoning. Our analysis reveals four recurring failure patterns: bare assertions, overlooked refutations, stagnation, and premature exit. Motivated by these findings, we examine whether explicit constraint-state tracking during execution mitigates these failures via LiveLedger, an inference-time tracker. This simple intervention consistently improves performance, substantially reducing underverified answers (by up to 26.5%) and improving overall accuracy (by up to 11.6%) on multi-constraint problems.
Authors: Kaleem Ullah Qasim, Jiashu Zhang, Hao Li, Muhammad Kafeel Shaheen
Abstract: Training language models to solve complex mathematical problems benefits from curriculum learning progressively training on simpler subproblems. However, existing decomposition methods are often heuristic, offering no guarantees that subproblems are simpler, that solving them aids the parent task, or that their relationships are mathematically grounded. We observe that symbolic differentiation provides a natural structure for verified decomposition: calculus rules explicitly define how expressions reduce to simpler components with provable properties. We introduce Verify-RL, a framework where every parent-child decomposition satisfies three verifiable conditions: strictly decreasing structural complexity, solution containment, and formal rule derivation. Unlike heuristic methods where a significant fraction of decompositions are invalid our properties admit automatic verification through symbolic computation, achieving "verification by construction" Experiments demonstrate that eliminating invalid decompositions yields sizable gains, accuracy on the hardest problems more than doubles from 32% to 68%, with a 40% relative improvement overall.
Authors: Junyu Feng, Binxiao Xu, Jiayi Chen, Mengyu Dai, Cenyang Wu, Haodong Li, Bohan Zeng, Yunliu Xie, Hao Liang, Ming Lu, Wentao Zhang
Abstract: This work addresses the challenge of personalized question answering in long-term human-machine interactions: when conversational history spans weeks or months and exceeds the context window, existing personalization mechanisms struggle to continuously absorb and leverage users' incremental concepts, aliases, and preferences. Current personalized multimodal models are predominantly static-concepts are fixed at initialization and cannot evolve during interactions. We propose M2A, an agentic dual-layer hybrid memory system that maintains personalized multimodal information through online updates. The system employs two collaborative agents: ChatAgent manages user interactions and autonomously decides when to query or update memory, while MemoryManager breaks down memory requests from ChatAgent into detailed operations on the dual-layer memory bank, which couples a RawMessageStore (immutable conversation log) with a SemanticMemoryStore (high-level observations), providing memories at different granularities. In addition, we develop a reusable data synthesis pipeline that injects concept-grounded sessions from Yo'LLaVA and MC-LLaVA into LoCoMo long conversations while preserving temporal coherence. Experiments show that M2A significantly outperforms baselines, demonstrating that transforming personalization from one-shot configuration to a co-evolving memory mechanism provides a viable path for high-quality individualized responses in long-term multimodal interactions. The code is available at https://github.com/Little-Fridge/M2A.
Authors: Keondo Park, Younghoon Na, Yourim Choi, Hyunwoo Ryu, Hyun-Woo Shin, Hyung-Sin Kim
Abstract: While the shift toward unified foundation models has revolutionized many deep learning domains, sleep medicine remains largely restricted to task-specific models that focus on localized micro-structure features. These approaches often neglect the rich, multi-modal context of Polysomnography (PSG) and fail to capture the global macro-structure of a full night's sleep. To address this, we introduce SleepMaMi , a Sleep Foundation Model engineered to master both hour-long sleep architectures and fine-grained signal morphologies. Our framework utilizes a hierarchical dual-encoder design: a Macro-Encoder to model full-night temporal dependencies and a Micro-Encoder to capture short-term characteristics from biosignals. Macro-Encoder is trained via Demographic-Guided Contrastive Learning, which aligns overnight sleep patterns with objective subject metadata, such as age, sex and BMI to refine global representations. Micro-Encoder is optimized via a hybrid Masked Autoencoder (MAE) and multi-modal contrastive objective. Pre-trained on a massive corpus of $>$20,000 PSG recordings (158K hours),SleepMaMi outperforms existing foundation models across a diverse suite of downstream tasks, demonstrating superior generalizability and label-efficient adaptation for clinical sleep analysis.
Authors: Zhuoyan Xu, Haoyang Fang, Boran Han, Bonan Min, Bernie Wang, Cuixiong Hu, Shuai Zhang
Abstract: Tabular data is frequently captured in image form across a wide range of real-world scenarios such as financial reports, handwritten records, and document scans. These visual representations pose unique challenges for machine understanding, as they combine both structural and visual complexities. While recent advances in Multimodal Large Language Models (MLLMs) show promising results in table understanding, they typically assume the relevant table is readily available. However, a more practical scenario involves identifying and reasoning over relevant tables from large-scale collections to answer user queries. To address this gap, we propose TabRAG, a framework that enables MLLMs to answer queries over large collections of table images. Our approach first retrieves candidate tables using jointly trained visual-text foundation models, then leverages MLLMs to perform fine-grained reranking of these candidates, and finally employs MLLMs to reason over the selected tables for answer generation. Through extensive experiments on a newly constructed dataset comprising 88,161 training and 9,819 testing samples across 8 benchmarks with 48,504 unique tables, we demonstrate that our framework significantly outperforms existing methods by 7.0% in retrieval recall and 6.1% in answer accuracy, offering a practical solution for real-world table understanding tasks.
Authors: Glenda Amaral, Tiago Prince Sales, Riccardo Baratella, Daniele Porello, Renata Guizzardi, Giancarlo Guizzardi
Abstract: Trust has stood out more than ever in the light of recent innovations. Some examples are advances in artificial intelligence that make machines more and more humanlike, and the introduction of decentralized technologies (e.g. blockchains), which creates new forms of (decentralized) trust. These new developments have the potential to improve the provision of products and services, as well as to contribute to individual and collective well-being. However, their adoption depends largely on trust. In order to build trustworthy systems, along with defining laws, regulations and proper governance models for new forms of trust, it is necessary to properly conceptualize trust, so that it can be understood both by humans and machines. This paper is the culmination of a long-term research program of providing a solid ontological foundation on trust, by creating reference conceptual models to support information modeling, automated reasoning, information integration and semantic interoperability tasks. To address this, a Reference Ontology of Trust (ONTrust) was developed, grounded on the Unified Foundational Ontology and specified in OntoUML, which has been applied in several initiatives, to demonstrate, for example, how it can be used for conceptual modeling and enterprise architecture design, for language evaluation and (re)design, for trust management, for requirements engineering, and for trustworthy artificial intelligence (AI) in the context of affective Human-AI teaming. ONTrust formally characterizes the concept of trust and its different types, describes the different factors that can influence trust, as well as explains how risk emerges from trust relations. To illustrate the working of ONTrust, the ontology is applied to model two case studies extracted from the literature.
Authors: Congcong Hu, Yuang Shi, Fan Huang, Yang Xiang, Zhou Ye, Ming Jin, Shiyu Wang
Abstract: Demand forecasting is a cornerstone of e-commerce operations, directly impacting inventory planning and fulfillment scheduling. However, existing forecasting systems often fail during high-impact periods such as flash sales, holiday campaigns, and sudden policy interventions, where demand patterns shift abruptly and unpredictably. In this paper, we introduce EventCast, a modular forecasting framework that integrates future event knowledge into time-series prediction. Unlike prior approaches that ignore future interventions or directly use large language models (LLMs) for numerical forecasting, EventCast leverages LLMs solely for event-driven reasoning. Unstructured business data, which covers campaigns, holiday schedules, and seller incentives, from existing operational databases, is processed by an LLM that converts it into interpretable textual summaries leveraging world knowledge for cultural nuances and novel event combinations. These summaries are fused with historical demand features within a dual-tower architecture, enabling accurate, explainable, and scalable forecasts. Deployed on real-world e-commerce scenarios spanning 4 countries of 160 regions over 10 months, EventCast achieves up to 86.9% and 97.7% improvement on MAE and MSE compared to the variant without event knowledge, and reduces MAE by up to 57.0% and MSE by 83.3% versus the best industrial baseline during event-driven periods. EventCast has deployed into real-world industrial pipelines since March 2025, offering a practical solution for improving operational decision-making in dynamic e-commerce environments.
Authors: Zhenyu Wu, Yanxi Long, Jian Li, Hua Huang
Abstract: Program code serves as a bridge linking vision and logic, providing a feasible supervisory approach for enhancing the multimodal reasoning capability of large models through geometric operations such as auxiliary line construction and perspective transformation. Nevertheless, current inverse graphics methods face tremendous challenges in accurately reconstructing complex geometric details, which often results in the loss of key geometric constraints or structural distortion. To address this bottleneck, we propose Geo-coder -- the first inverse programming framework for geometric images based on a multi-agent system. Our method innovatively decouples the process into geometric modeling via pixel-wise anchoring and metric-driven code evolution: Stage 1 leverages the complementary advantages of visual operators and large models to achieve precise capture of pixel coordinates and visual attributes; Stage 2 introduces a synthesis-rendering-validation closed loop, where bidirectional visual feedback drives the self-correction of code. Extensive experiments demonstrate that Geo-coder achieves a substantial lead in both geometric reconstruction accuracy and visual consistency. Notably, by effectively preserving the core geometric semantics, the images reconstructed with our method exhibit equivalent performance to the original ones in multimodal reasoning tasks, which fully validates the robustness of the framework. Finally, to further reduce research costs, we have open-sourced the Geo-coder dataset constructed on the GeoCode framework, which contains more than 1,500 samples. On this basis, we have also open-sourced the GeocodeLM model, laying a solid data and model foundation for subsequent research in this field.
Authors: Bahare Riahi, Veronica Catete
Abstract: This study investigates students' perceptions of Artificial Intelligence (AI) grading systems in an undergraduate computer science course (n = 27), focusing on a block-based programming final project. Guided by the ethical principles framework articulated by Jobin (2019), our study examines fairness, trust, consistency, and transparency in AI grading by comparing AI-generated feedback with original human-graded feedback. Findings reveal concerns about AI's lack of contextual understanding and personalization. We recommend that equitable and trustworthy AI systems reflect human judgment, flexibility, and empathy, serving as supplementary tools under human oversight. This work contributes to ethics-centered assessment practices by amplifying student voices and offering design principles for humanizing AI in designed learning environments.
Authors: Yiming Xiong, Shengran Hu, Jeff Clune
Abstract: The statelessness of foundation models bottlenecks agentic systems' ability to continually learn, a core capability for long-horizon reasoning and adaptation. To address this limitation, agentic systems commonly incorporate memory modules to retain and reuse past experience, aiming for continual learning during test time. However, most existing memory designs are human-crafted and fixed, which limits their ability to adapt to the diversity and non-stationarity of real-world tasks. In this paper, we introduce ALMA (Automated meta-Learning of Memory designs for Agentic systems), a framework that meta-learns memory designs to replace hand-engineered memory designs, therefore minimizing human effort and enabling agentic systems to be continual learners across diverse domains. Our approach employs a Meta Agent that searches over memory designs expressed as executable code in an open-ended manner, theoretically allowing the discovery of arbitrary memory designs, including database schemas as well as their retrieval and update mechanisms. Extensive experiments across four sequential decision-making domains demonstrate that the learned memory designs enable more effective and efficient learning from experience than state-of-the-art human-crafted memory designs on all benchmarks. When developed and deployed safely, ALMA represents a step toward self-improving AI systems that learn to be adaptive, continual learners.
Authors: Zhirong Huang, Debo Cheng, Guixian Zhang, Yi Wang, Jiuyong Li, Shichao Zhang
Abstract: Instrumental variables (IVs) are crucial for addressing unobservable confounders, yet their stringent exogeneity assumptions pose significant challenges in networked data. Existing methods typically rely on modelling neighbour information when recovering IVs, thereby inevitably mixing shared environment-induced endogenous correlations and individual-specific exogenous variation, leading the resulting IVs to inherit dependence on unobserved confounders and to violate exogeneity. To overcome this challenge, we propose $\underline{Dis}$entangled $\underline{I}$nstrumental $\underline{V}$ariables (DisIV) framework, a novel method for causal inference based on networked observational data with latent confounders. DisIV exploits network homogeneity as an inductive bias and employs a structural disentanglement mechanism to extract individual-specific components that serve as latent IVs. The causal validity of the extracted IVs is constrained through explicit orthogonality and exclusion conditions. Extensive semi-synthetic experiments on real-world datasets demonstrate that DisIV consistently outperforms state-of-the-art baselines in causal effect estimation under network-induced confounding.
Authors: Pierre-Louis Favreau, Jean-Pierre Lo, Clement Guiguet, Charles Simon-Meunier, Nicolas Dehandschoewercker, Allen G. Roush, Judah Goldfeder, Ravid Shwartz-Ziv
Abstract: We present Minitap, a multi-agent system that achieves 100% success on the AndroidWorld benchmark, the first to fully solve all 116 tasks and surpassing human performance (80%). We first analyze why single-agent architectures fail: context pollution from mixed reasoning traces, silent text input failures undetected by the agent, and repetitive action loops without escape. Minitap addresses each failure through targeted mechanisms: cognitive separation across six specialized agents, deterministic post-validation of text input against device state, and meta-cognitive reasoning that detects cycles and triggers strategy changes. Ablations show multi-agent decomposition contributes +21 points over single-agent baselines; verified execution adds +7 points; meta-cognition adds +9 points. We release Minitap as open-source software. https://github.com/minitap-ai/mobile-use
Authors: Yiwei Qin, Zhen Huang, Tiantian Mi, Weiye Si, Chenyang Zhou, Qipeng Guo, Siyuan Feng, Pengfei Liu
Abstract: Data quality determines foundation model performance, yet systematic processing frameworks are lacking. We introduce Data Darwinism, a ten-level taxonomy (L0-L9) that conceptualizes data-model co-evolution: advanced models produce superior data for next-generation systems. We validate this on scientific literature by constructing Darwin-Science, a 900B-token corpus (L0-L5). We identify a learnability gap in raw scientific text, which we bridge via L4 (Generative Refinement) and L5 (Cognitive Completion) using frontier LLMs to explicate reasoning and terminology. To ensure rigorous attribution, we pre-trained daVinci-origin-3B/7B models from scratch, excluding scientific content to create contamination-free baselines. After 600B tokens of continued pre-training, Darwin-Science outperforms baselines by +2.12 (3B) and +2.95 (7B) points across 20+ benchmarks, rising to +5.60 and +8.40 points on domain-aligned tasks. Systematic progression to L5 yields a +1.36 total gain, confirming that higher-level processing unlocks latent data value. We release the Darwin-Science corpus and daVinci-origin models to enable principled, co-evolutionary development.
Authors: Jiahui Zhou, Dan Li, Boxin Li, Xiao Zhang, Erli Meng, Lin Li, Zhuomin Chen, Jian Lou, See-Kiong Ng
Abstract: Time series is a pervasive data type across various application domains, rendering the reasonable solving of diverse time series tasks a long-standing goal. Recent advances in large language models (LLMs), especially their reasoning abilities unlocked through reinforcement learning (RL), have opened new opportunities for tackling tasks with long Chain-of-Thought (CoT) reasoning. However, leveraging LLM reasoning for time series remains in its infancy, hindered by the absence of carefully curated time series CoT data for training, limited data efficiency caused by underexplored data scheduling, and the lack of RL algorithms tailored for exploiting such time series CoT data. In this paper, we introduce VeriTime, a framework that tailors LLMs for time series reasoning through data synthesis, data scheduling, and RL training. First, we propose a data synthesis pipeline that constructs a TS-text multimodal dataset with process-verifiable annotations. Second, we design a data scheduling mechanism that arranges training samples according to a principled hierarchy of difficulty and task taxonomy. Third, we develop a two-stage reinforcement finetuning featuring fine-grained, multi-objective rewards that leverage verifiable process-level CoT data. Extensive experiments show that VeriTime substantially boosts LLM performance across diverse time series reasoning tasks. Notably, it enables compact 3B, 4B models to achieve reasoning capabilities on par with or exceeding those of larger proprietary LLMs.
Authors: Xin Wang, Hualin Zhou, Sheng Guang Wang, Ting Dang, Yu Zhang, Hong Jia, Tao Gu
Abstract: Deploying Vision-Language Models (VLMs) on edge devices is challenged by resource constraints and performance degradation under distribution shifts. While test-time adaptation (TTA) can counteract such shifts, existing methods are too resource-intensive for on-device deployment. To address this challenge, we propose LQA, a lightweight, quantized-adaptive framework for VLMs that combines a modality-aware quantization strategy with gradient-free test-time adaptation. We introduce Selective Hybrid Quantization (SHQ) and a quantized, gradient-free adaptation mechanism to enable robust and efficient VLM deployment on resource-constrained hardware. Experiments across both synthetic and real-world distribution shifts show that LQA improves overall adaptation performance by 4.5\%, uses less memory than full-precision models, and significantly outperforms gradient-based TTA methods, achieving up to 19.9$\times$ lower memory usage across seven open-source datasets. These results demonstrate that LQA offers a practical pathway for robust, privacy-preserving, and efficient VLM deployment on edge devices.
Authors: Anna Soligo, Edward Turner, Senthooran Rajamanoharan, Neel Nanda
Abstract: Finetuning large language models on narrowly harmful datasets can cause them to become emergently misaligned, giving stereotypically `evil' responses across diverse unrelated settings. Concerningly, a pre-registered survey of experts failed to predict this result, highlighting our poor understanding of the inductive biases governing learning and generalisation in LLMs. We use emergent misalignment (EM) as a case study to investigate these inductive biases and find that models can just learn the narrow dataset task, but that the general solution appears to be more stable and more efficient. To establish this, we build on the result that different EM finetunes converge to the same linear representation of general misalignment, which can be used to mediate misaligned behaviour. We find a linear representation of the narrow solution also exists, and can be learned by introducing a KL divergence loss. Comparing these representations reveals that general misalignment achieves lower loss, is more robust to perturbations, and is more influential in the pre-training distribution. This work isolates a concrete representation of general misalignment for monitoring and mitigation. More broadly, it offers a detailed case study and preliminary metrics for investigating how inductive biases shape generalisation in LLMs. We open-source all code, datasets and model finetunes.
Authors: Jingqi Zhou, Sheng Wang, DeZhao Deng, Junwen Lu, Junwei Su, Qintong Li, Jiahui Gao, Hao Wu, Jiyue Jiang, Lingpeng Kong, Chuan Wu
Abstract: Agentic systems powered by Large Language Models (LLMs) have demonstrated remarkable potential in tackling complex, long-horizon tasks. However, their efficacy is fundamentally constrained by static configurations governing agent behaviors, which are fixed prior to execution and fail to adapt to evolving task dynamics. Existing approaches, relying on manual orchestration or heuristic-based patches, often struggle with poor generalization and fragmented optimization. To transcend these limitations, we propose ToolSelf, a novel paradigm enabling tool-driven runtime self-reconfiguration. By abstracting configuration updates as a callable tool, ToolSelf unifies task execution and self-adjustment into a single action space, achieving a phase transition from external rules to intrinsic parameters. Agents can thereby autonomously update their sub-goals and context based on task progression, and correspondingly adapt their strategy and toolbox, transforming from passive executors into dual managers of both task and self. We further devise Configuration-Aware Two-stage Training (CAT), combining rejection sampling fine-tuning with trajectory-level reinforcement learning to internalize this meta-capability. Extensive experiments across diverse benchmarks demonstrate that ToolSelf rivals specialized workflows while generalizing to novel tasks, achieving a 24.1% average performance gain and illuminating a path toward truly self-adaptive agents.
Authors: Zhenyuan Zhang, Xianzhang Jia, Zhiqin Yang, Zhenbo Song, Wei Xue, Sirui Han, Yike Guo
Abstract: Long-term memory enables large language model agents to tackle complex tasks through historical interactions. However, existing frameworks encounter a fundamental dilemma between compressing redundant information efficiently and maintaining precise retrieval for downstream tasks. To bridge this gap, we propose MemFly, a framework grounded in information bottleneck principles that facilitates on-the-fly memory evolution for LLMs. Our approach minimizes compression entropy while maximizing relevance entropy via a gradient-free optimizer, constructing a stratified memory structure for efficient storage. To fully leverage MemFly, we develop a hybrid retrieval mechanism that seamlessly integrates semantic, symbolic, and topological pathways, incorporating iterative refinement to handle complex multi-hop queries. Comprehensive experiments demonstrate that MemFly substantially outperforms state-of-the-art baselines in memory coherence, response fidelity, and accuracy.
Authors: Mingcan Wang, Junchang Xin, Zhongming Yao, Kaifu Long, Zhiqiong Wang
Abstract: The algorithms based on message passing neural networks (MPNNs) on graphs have recently achieved great success for various graph applications. However, studies find that these methods always propagate the information to very limited neighborhoods with shallow depth, particularly due to over-smoothing. That means most of the existing MPNNs fail to be so `deep'. Although some previous work tended to handle this challenge via optimization- or structure-level remedies, the overall performance of GCNs still suffers from limited accuracy, poor stability, and unaffordable computational cost. Moreover, neglect of higher-order relationships during the propagation of MPNNs has further limited the performance of them. To overcome these challenges, a novel variant of PageRank named motif-based personalized PageRank (MPPR) is proposed to measure the influence of one node to another on the basis of considering higher-order motif relationships. Secondly, the MPPR is utilized to the message passing process of GCNs, thereby guiding the message passing process at a relatively `high' level. The experimental results show that the proposed method outperforms almost all of the baselines on accuracy, stability, and time consumption. Additionally, the proposed method can be considered as a component that can underpin almost all GCN tasks, with DGCRL being demonstrated in the experiment. The anonymous code repository is available at: https://anonymous.4open.science/r/GCN-MPPR-AFD6/.
Authors: Yu Zhao, Hao Guan, Yongcheng Jing, Ying Zhang, Dacheng Tao
Abstract: Large Language Models (LLMs) have shown strong potential in complex medical reasoning yet face diminishing gains under inference scaling laws. While existing studies augment LLMs with various knowledge types, it remains unclear how effectively the additional costs translate into accuracy. In this paper, we explore how meta-cognition of LLMs, i.e., their self-awareness of their own knowledge states, can regulate the reasoning process. Specifically, we propose MedCoG, a Medical Meta-Cognition Agent with Knowledge Graph, where the meta-cognitive assessments of task complexity, familiarity, and knowledge density dynamically regulate utilization of procedural, episodic, and factual knowledge. The LLM-centric on-demand reasoning aims to mitigate scaling laws by (1) reducing costs via avoiding indiscriminate scaling, (2) improving accuracy via filtering out distractive knowledge. To validate this, we empirically characterize the scaling curve and introduce inference density to quantify inference efficiency, defined as the ratio of theoretically effective cost to actual cost. Experiments demonstrate the effectiveness and efficiency of MedCoG on five hard sets of medical benchmarks, yielding 5.5x inference density. Furthermore, the Oracle study highlights the significant potential of meta-cognitive regulation.
Authors: Mansi, Avinash Kori, Francesca Toni, Soteris Demetriou
Abstract: Text guided diffusion models are used by millions of users, but can be easily exploited to produce harmful content. Concept unlearning methods aim at reducing the models' likelihood of generating harmful content. Traditionally, this has been tackled at an individual concept level, with only a handful of recent works considering more realistic concept combinations. However, state of the art methods depend on full finetuning, which is computationally expensive. Concept localisation methods can facilitate selective finetuning, but existing techniques are static, resulting in suboptimal utility. In order to tackle these challenges, we propose TRUST (Targeted Robust Selective fine Tuning), a novel approach for dynamically estimating target concept neurons and unlearning them through selective finetuning, empowered by a Hessian based regularization. We show experimentally, against a number of SOTA baselines, that TRUST is robust against adversarial prompts, preserves generation quality to a significant degree, and is also significantly faster than the SOTA. Our method achieves unlearning of not only individual concepts but also combinations of concepts and conditional concepts, without any specific regularization.
Authors: Guanglong Sun, Hongwei Yan, Liyuan Wang, Zhiqi Kang, Shuang Cui, Hang Su, Jun Zhu, Yi Zhong
Abstract: To cope with uncertain changes of the external world, intelligent systems must continually learn from complex, evolving environments and respond in real time. This ability, collectively known as general continual learning (GCL), encapsulates practical challenges such as online datastreams and blurry task boundaries. Although leveraging pretrained models (PTMs) has greatly advanced conventional continual learning (CL), these methods remain limited in reconciling the diverse and temporally mixed information along a single pass, resulting in sub-optimal GCL performance. Inspired by meta-plasticity and reconstructive memory in neuroscience, we introduce here an innovative approach named Meta Post-Refinement (MePo) for PTMs-based GCL. This approach constructs pseudo task sequences from pretraining data and develops a bi-level meta-learning paradigm to refine the pretrained backbone, which serves as a prolonged pretraining phase but greatly facilitates rapid adaptation of representation learning to downstream GCL tasks. MePo further initializes a meta covariance matrix as the reference geometry of pretrained representation space, enabling GCL to exploit second-order statistics for robust output alignment. MePo serves as a plug-in strategy that achieves significant performance gains across a variety of GCL benchmarks and pretrained checkpoints in a rehearsal-free manner (e.g., 15.10\%, 13.36\%, and 12.56\% on CIFAR-100, ImageNet-R, and CUB-200 under Sup-21/1K). Our source code is available at \href{https://github.com/SunGL001/MePo}{MePo}
Authors: Ivaxi Sheth, Zhijing Jin, Bryan Wilder, Dominik Janzing, Mario Fritz
Abstract: In the presence of confounding between an endogenous variable and the outcome, instrumental variables (IVs) are used to isolate the causal effect of the endogenous variable. Identifying valid instruments requires interdisciplinary knowledge, creativity, and contextual understanding, making it a non-trivial task. In this paper, we investigate whether large language models (LLMs) can aid in this task. We perform a two-stage evaluation framework. First, we test whether LLMs can recover well-established instruments from the literature, assessing their ability to replicate standard reasoning. Second, we evaluate whether LLMs can identify and avoid instruments that have been empirically or theoretically discredited. Building on these results, we introduce IV Co-Scientist, a multi-agent system that proposes, critiques, and refines IVs for a given treatment-outcome pair. We also introduce a statistical test to contextualize consistency in the absence of ground truth. Our results show the potential of LLMs to discover valid instrumental variables from a large observational database.
Authors: Weihao Zeng, Yuzhen Huang, Junxian He
Abstract: Large language models (LLMs) are increasingly capable of carrying out long-running, real-world tasks. However, as the amount of context grows, their reliability often deteriorates, a phenomenon known as "context rot". Existing long-context benchmarks primarily focus on single-step settings that evaluate a model's ability to retrieve information from a long snippet. In realistic scenarios, however, LLMs often need to act as agents that explore environments, follow instructions and plans, extract useful information, and predict correct actions under a dynamically growing context. To assess language agents in such settings, we introduce LOCA-bench (a benchmark for LOng-Context Agents). Given a task prompt, LOCA-bench leverages automated and scalable control of environment states to regulate the agent's context length. This design enables LOCA-bench to extend the context length potentially to infinity in a controlled way while keeping the underlying task semantics fixed. LOCA-bench evaluates language agents as a combination of models and scaffolds, including various context management strategies. While agent performance generally degrades as the environment states grow more complex, advanced context management techniques can substantially improve the overall success rate. We open-source LOCA-bench to provide a platform for evaluating models and scaffolds in long-context, agentic scenarios: https://github.com/hkust-nlp/LOCA-bench
Authors: Jishu Sen Gupta, Harini SI, Somesh Kumar Singh, Syed Mohamad Tawseeq, Yaman Kumar Singla, David Doermann, Rajiv Ratn Shah, Balaji Krishnamurthy
Abstract: Data-driven social science research is inherently slow, relying on iterative cycles of observation, hypothesis generation, and experimental validation. While recent data-driven methods promise to accelerate parts of this process, they largely fail to support end-to-end scientific discovery. To address this gap, we introduce EXPERIGEN, an agentic framework that operationalizes end-to-end discovery through a Bayesian optimization inspired two-phase search, in which a Generator proposes candidate hypotheses and an Experimenter evaluates them empirically. Across multiple domains, EXPERIGEN consistently discovers 2-4x more statistically significant hypotheses that are 7-17 percent more predictive than prior approaches, and naturally extends to complex data regimes including multimodal and relational datasets. Beyond statistical performance, hypotheses must be novel, empirically grounded, and actionable to drive real scientific progress. To evaluate these qualities, we conduct an expert review of machine-generated hypotheses, collecting feedback from senior faculty. Among 25 reviewed hypotheses, 88 percent were rated moderately or strongly novel, 70 percent were deemed impactful and worth pursuing, and most demonstrated rigor comparable to senior graduate-level research. Finally, recognizing that ultimate validation requires real-world evidence, we conduct the first A/B test of LLM-generated hypotheses, observing statistically significant results with p less than 1e-6 and a large effect size of 344 percent.
Authors: Rui Li, Zeyu Zhang, Xiaohe Bo, Quanyu Dai, Chaozhuo Li, Feng Wen, Xu Chen
Abstract: Multi-agent architectures built on large language models (LLMs) have demonstrated the potential to realize swarm intelligence through well-crafted collaboration. However, the substantial burden of manual orchestration inherently raises an imperative to automate the design of agentic workflows. We frame such an agent coordination challenge as a classic problem in dynamic ad-hoc networking: How to establish adaptive and reliable communication among a scalable number of agentic hosts? In response to this unresolved dilemma, we introduce RAPS, a reputation-aware publish-subscribe paradigm for adaptive, scalable, and robust coordination of LLM agents. RAPS is grounded in the Distributed Publish-Subscribe Protocol, allowing LLM agents to exchange messages based on their declared intents rather than predefined topologies. Beyond this substrate, RAPS further incorporates two coherent overlays: (i) Reactive Subscription, enabling agents to dynamically refine their intents; and (ii) Bayesian Reputation, empowering each agent with a local watchdog to detect and isolate malicious peers. Extensive experiments over five benchmarks showcase that our design effectively reconciles adaptivity, scalability, and robustness in a unified multi-agent coordination framework.
Authors: Yuqiao Meng, Luoxi Tang, Dazheng Zhang, Rafael Brens, Elvys J. Romero, Nancy Guo, Safa Elkefi, Zhaohan Xi
Abstract: The rapid adoption of large language models (LLMs) in digital health has been driven by a "scaling-first" philosophy, i.e., the assumption that clinical intelligence increases with model size and data. However, real-world clinical needs include not only effectiveness, but also reliability and reasonable deployment cost. Since clinical decision-making is inherently collaborative, we challenge the monolithic scaling paradigm and ask whether a Small Agent Group (SAG) can support better clinical reasoning. SAG shifts from single-model intelligence to collective expertise by distributing reasoning, evidence-based analysis, and critical audit through a collaborative deliberation process. To assess the clinical utility of SAG, we conduct extensive evaluations using diverse clinical metrics spanning effectiveness, reliability, and deployment cost. Our results show that SAG achieves superior performance compared to a single giant model, both with and without additional optimization or retrieval-augmented generation. These findings suggest that the synergistic reasoning represented by SAG can substitute for model parameter growth in clinical settings. Overall, SAG offers a scalable solution to digital health that better balances effectiveness, reliability, and deployment efficiency.
Authors: Zhan-Yi Liao, Jaewon Yoo, Hao-Tsung Yang, Po-An Chen
Abstract: Counterfactual explanation (CE) is a core technique in explainable artificial intelligence (XAI), widely used to interpret model decisions and suggest actionable alternatives. This work presents a structure-aware and robustness-oriented counterfactual search method based on the conditional Gaussian network classifier (CGNC). The CGNC has a generative structure that encodes conditional dependencies and potential causal relations among features through a directed acyclic graph (DAG). This structure naturally embeds feature relationships into the search process, eliminating the need for additional constraints to ensure consistency with the model's structural assumptions. We adopt a convergence-guaranteed cutting-set procedure as an adversarial optimization framework, which iteratively approximates solutions that satisfy global robustness conditions. To address the nonconvex quadratic structure induced by feature dependencies, we apply piecewise McCormick relaxation to reformulate the problem as a mixed-integer linear program (MILP), ensuring global optimality. Experimental results show that our method achieves strong robustness, with direct global optimization of the original formulation providing especially stable and efficient results. The proposed framework is extensible to more complex constraint settings, laying the groundwork for future advances in counterfactual reasoning under nonconvex quadratic formulations.
Authors: Yilun Zheng, Dongyang Ma, Tian Liang, Jiahao Xu, Xinting Huang, Lijie Chen, Haitao Mi, Yan Wang
Abstract: Reasoning models enhance problem-solving by scaling test-time compute, yet they face a critical paradox: excessive thinking tokens often degrade performance rather than improve it. We attribute this to a fundamental architectural flaw: standard LLMs operate as "malloc-only" engines, continuously accumulating valid and redundant steps alike without a mechanism to prune obsolete information. To break this cycle, we propose Free()LM, a model that introduces an intrinsic self-forgetting capability via the Free-Module, a plug-and-play LoRA adapter. By iteratively switching between reasoning and cleaning modes, Free()LM dynamically identifies and prunes useless context chunks, maintaining a compact and noise-free state. Extensive experiments show that Free()LM provides consistent improvements across all model scales (8B to 685B). It achieves a 3.3% average improvement over top-tier reasoning baselines, even establishing a new SOTA on IMOanswerBench using DeepSeek V3.2-Speciale. Most notably, in long-horizon tasks where the standard Qwen3-235B-A22B model suffers a total collapse (0% accuracy), Free()LM restores performance to 50%. Our findings suggest that sustainable intelligence requires the freedom to forget as much as the power to think.
Authors: Bulent Soykan, Sean Mondesire, Ghaith Rabadi, Grace Bochenek
Abstract: The Unrelated Parallel Machine Scheduling Problem (UPMSP) with release dates, setups, and eligibility constraints presents a significant multi-objective challenge. Traditional methods struggle to balance minimizing Total Weighted Tardiness (TWT) and Total Setup Time (TST). This paper proposes a Deep Reinforcement Learning framework using Proximal Policy Optimization (PPO) and a Graph Neural Network (GNN). The GNN effectively represents the complex state of jobs, machines, and setups, allowing the PPO agent to learn a direct scheduling policy. Guided by a multi-objective reward function, the agent simultaneously minimizes TWT and TST. Experimental results on benchmark instances demonstrate that our PPO-GNN agent significantly outperforms a standard dispatching rule and a metaheuristic, achieving a superior trade-off between both objectives. This provides a robust and scalable solution for complex manufacturing scheduling.
Authors: Doni Bloomfield, Allison Berke, Moritz S. Hanke, Aaron Maiwald, James R. M. Black, Toby Webster, Tina Hernandez-Boussard, Oliver M. Crook, Jassi Pannu
Abstract: Training data is an essential input into creating competent artificial intelligence (AI) models. AI models for biology are trained on large volumes of data, including data related to biological sequences, structures, images, and functions. The type of data used to train a model is intimately tied to the capabilities it ultimately possesses--including those of biosecurity concern. For this reason, an international group of more than 100 researchers at the recent 50th anniversary Asilomar Conference endorsed data controls to prevent the use of AI for harmful applications such as bioweapons development. To help design such controls, we introduce a five-tier Biosecurity Data Level (BDL) framework for categorizing pathogen data. Each level contains specific data types, based on their expected ability to contribute to capabilities of concern when used to train AI models. For each BDL tier, we propose technical restrictions appropriate to its level of risk. Finally, we outline a novel governance framework for newly created dual-use pathogen data. In a world with widely accessible computational and coding resources, data controls may be among the most high-leverage interventions available to reduce the proliferation of concerning biological AI capabilities.
Authors: Majid Ghasemi, Mark Crowley
Abstract: Contemporary AI alignment strategies rely on a fragile premise: that human feedback, while noisy, remains a fundamentally truthful signal. In this paper, we identify this assumption as Dogma 4 of Reinforcement Learning (RL). We demonstrate that while this dogma holds in static environments, it fails in social settings where evaluators may be sycophantic, lazy, or adversarial. We prove that under Dogma 4, standard RL agents suffer from what we call Objective Decoupling, a structural failure mode where the agent's learned objective permanently separates from the latent ground truth, guaranteeing convergence to misalignment. To resolve this, we propose Epistemic Source Alignment (ESA). Unlike standard robust methods that rely on statistical consensus (trusting the majority), ESA utilizes sparse safety axioms to judge the source of the feedback rather than the signal itself. We prove that this "judging the judges" mechanism guarantees convergence to the true objective, even when a majority of evaluators are biased. Empirically, we show that while traditional consensus methods fail under majority collusion, our approach successfully recovers the optimal policy.
Authors: Risal Shahriar Shefin, Debashis Gupta, Thai Le, Sarra Alqahtani
Abstract: Multi-Agent Reinforcement Learning (MARL) is increasingly deployed in safety-critical domains, yet methods for interpretable failure detection and attribution remain underdeveloped. We introduce a two-stage gradient-based framework that provides interpretable diagnostics for three critical failure analysis tasks: (1) detecting the true initial failure source (Patient-0); (2) validating why non-attacked agents may be flagged first due to domino effects; and (3) tracing how failures propagate through learned coordination pathways. Stage 1 performs interpretable per-agent failure detection via Taylor-remainder analysis of policy-gradient costs, declaring an initial Patient-0 candidate at the first threshold crossing. Stage 2 provides validation through geometric analysis of critic derivatives-first-order sensitivity and directional second-order curvature aggregated over causal windows to construct interpretable contagion graphs. This approach explains "downstream-first" detection anomalies by revealing pathways that amplify upstream deviations. Evaluated across 500 episodes in Simple Spread (3 and 5 agents) and 100 episodes in StarCraft II using MADDPG and HATRPO, our method achieves 88.2-99.4% Patient-0 detection accuracy while providing interpretable geometric evidence for detection decisions. By moving beyond black-box detection to interpretable gradient-level forensics, this framework offers practical tools for diagnosing cascading failures in safety-critical MARL systems.
Authors: Liying Wang, Madison Lee, Yunzhang Jiang, Steven Chen, Kewei Sha, Yunhe Feng, Frank Wong, Lisa Hightow-Weidman, Weichao Yuwen
Abstract: Background: HIV and substance use represent interacting epidemics with shared psychological drivers - impulsivity and maladaptive coping. Dialectical behavior therapy (DBT) targets these mechanisms but faces scalability challenges. Generative artificial intelligence (GenAI) offers potential for delivering personalized DBT coaching at scale, yet rapid development has outpaced safety infrastructure. Methods: We developed Glow, a GenAI-powered DBT skills coach delivering chain and solution analysis for individuals at risk for HIV and substance use. In partnership with a Los Angeles community health organization, we conducted usability testing with clinical staff (n=6) and individuals with lived experience (n=28). Using the Helpful, Honest, and Harmless (HHH) framework, we employed user-driven adversarial testing wherein participants identified target behaviors and generated contextually realistic risk probes. We evaluated safety performance across 37 risk probe interactions. Results: Glow appropriately handled 73% of risk probes, but performance varied by agent. The solution analysis agent demonstrated 90% appropriate handling versus 44% for the chain analysis agent. Safety failures clustered around encouraging substance use and normalizing harmful behaviors. The chain analysis agent fell into an "empathy trap," providing validation that reinforced maladaptive beliefs. Additionally, 27 instances of DBT skill misinformation were identified. Conclusions: This study provides the first systematic safety evaluation of GenAI-delivered DBT coaching for HIV and substance use risk reduction. Findings reveal vulnerabilities requiring mitigation before clinical trials. The HHH framework and user-driven adversarial testing offer replicable methods for evaluating GenAI mental health interventions.
Authors: Ziwei Wang, Yuanhe Zhang, Jing Chen, Zhenhong Zhou, Ruichao Liang, Ruiying Du, Ju Jia, Cong Wu, Yang Liu
Abstract: Large Reasoning Models (LRMs) employ reasoning to address complex tasks. Such explicit reasoning requires extended context lengths, resulting in substantially higher resource consumption. Prior work has shown that adversarially crafted inputs can trigger redundant reasoning processes, exposing LRMs to resource-exhaustion vulnerabilities. However, the reasoning process itself, especially its reflective component, has received limited attention, even though it can lead to over-reflection and consume excessive computing power. In this paper, we introduce Recursive Entropy to quantify the risk of resource consumption in reflection, thereby revealing the safety issues inherent in inference itself. Based on Recursive Entropy, we introduce RECUR, a resource exhaustion attack via Recursive Entropy guided Counterfactual Utilization and Reflection. It constructs counterfactual questions to verify the inherent flaws and risks of LRMs. Extensive experiments demonstrate that, under benign inference, recursive entropy exhibits a pronounced decreasing trend. RECUR disrupts this trend, increasing the output length by up to 11x and decreasing throughput by 90%. Our work provides a new perspective on robust reasoning.
Authors: Zehao Chen, Gongxun Li, Tianxiang Ai, Yifei Li, Zixuan Huang, Wang Zhou, Fuzhen Zhuang, Xianglong Liu, Jianxin Li, Deqing Wang, Yikun Ban
Abstract: As post-training optimization becomes central to improving large language models, we observe a persistent saturation bottleneck: once models grow highly confident, further training yields diminishing returns. While existing methods continue to reinforce target predictions, we find that informative supervision signals remain latent in models' own historical weak states. Motivated by this observation, we propose WMSS (Weak Agents Can Make Strong Agents Stronger), a post-training paradigm that leverages weak checkpoints to guide continued optimization. By identifying recoverable learning gaps via entropy dynamics and reinforcing them through compensatory learning, WMSS enables strong agents to improve beyond conventional post-training saturation. Experiments on mathematical reasoning and code generation datasets show that agents trained with our approach achieve effective performance improvements, while incurring zero additional inference cost.
Authors: Yifan Yang, Jinjia Li, Kunxi Li, Puhao Zheng, Yuanyi Wang, Zheyan Qu, Yang Yu, Jianmin Wu, Ming Li, Hongxia Yang
Abstract: The rapid advancement of large language models (LLMs) demands increasingly reliable evaluation, yet current centralized evaluation suffers from opacity, overfitting, and hardware-induced variance. Our empirical analysis reveals an alarming inconsistency in existing evaluations: the standard deviation across ten repeated runs of a single model on HumanEval (1.67) actually exceeds the performance gap among the top-10 models on the official leaderboard (0.91), rendering current rankings statistically precarious. To mitigate these instabilities, we propose a decentralized evaluation framework that enables hardware and parameter diversity through large-scale benchmarking across heterogeneous compute nodes. By leveraging the blockchain-based protocol, the framework incentivizes global contributors to act as independent validators, using a robust reward system to ensure evaluation integrity and discourage dishonest participation. This collective verification transforms evaluation from a "centralized black box" into a "decentralized endorsement" where multi-party consensus and diverse inference environments yield a more stable, representative metric. Experimental results demonstrate that the decentralized evaluation framework reduces the standard deviation across ten runs on the same model to 0.28. This significant improvement over conventional frameworks ensures higher statistical confidence in model rankings. We have completely implemented this platform and will soon release it to the community.
Authors: Xun Su, Huamin Wang, Qi Zhang
Abstract: Speech Emotion Recognition (SER) is widely deployed in Human-Computer Interaction, yet the high computational cost of conventional models hinders their implementation on resource-constrained edge devices. Spiking Neural Networks (SNNs) offer an energy-efficient alternative due to their event-driven nature; however, their integration with continuous Self-Supervised Learning (SSL) representations is fundamentally challenged by distribution mismatch, where high-dynamic-range embeddings degrade the information coding capacity of threshold-based neurons. To resolve this, we propose Prompt-Tuned Spiking Neural Networks (PTS-SNN), a parameter-efficient neuromorphic adaptation framework that aligns frozen SSL backbones with spiking dynamics. Specifically, we introduce a Temporal Shift Spiking Encoder to capture local temporal dependencies via parameter-free channel shifts, establishing a stable feature basis. To bridge the domain gap, we devise a Context-Aware Membrane Potential Calibration strategy. This mechanism leverages a Spiking Sparse Linear Attention module to aggregate global semantic context into learnable soft prompts, which dynamically regulate the bias voltages of Parametric Leaky Integrate-and-Fire (PLIF) neurons. This regulation effectively centers the heterogeneous input distribution within the responsive firing range, mitigating functional silence or saturation. Extensive experiments on five multilingual datasets (e.g., IEMOCAP, CASIA, EMODB) demonstrate that PTS-SNN achieves 73.34\% accuracy on IEMOCAP, comparable to competitive Artificial Neural Networks (ANNs), while requiring only 1.19M trainable parameters and 0.35 mJ inference energy per sample.
Authors: Siqu Ou, Tianrui Wan, Zhiyuan Zhao, Junyu Gao, Xuelong Li
Abstract: While chain-of-thought (CoT) reasoning has substantially improved multimodal large language models (MLLMs) on complex reasoning tasks, existing approaches largely rely on long textual reasoning trajectories and provide limited mechanisms for learning stable visual attention policies. Our analysis shows that current MLLMs exhibit weak visual focus: early-stage visual misalignment is rarely corrected during subsequent reasoning, leading to error propagation and failed inferences. We argue that this limitation stems from inadequate credit assignment for visual attention during training. To address this issue, we propose SAYO, a visual reasoning model trained with a reinforcement learning (RL) framework that introduces a region-level visual attention-based reward. This reward explicitly aligns optimization signals with visually grounded reasoning steps, enabling the model to learn more reliable attention behaviors. Extensive experiments across multiple multimodal benchmarks demonstrate that SAYO consistently improves performance on diverse reasoning and perception tasks.
Authors: Baoyun Zhao, He Wang, Liang Zeng
Abstract: While Large Language Models (LLMs) have recently shown promise in Automated Heuristic Design (AHD), existing approaches typically formulate AHD around constructive priority rules or parameterized local search guidance, thereby restricting the search space to fixed heuristic forms. Such designs offer limited capacity for structural exploration, making it difficult to escape deep local optima in complex Combinatorial Optimization Problems (COPs). In this work, we propose G-LNS, a generative evolutionary framework that extends LLM-based AHD to the automated design of Large Neighborhood Search (LNS) operators. Unlike prior methods that evolve heuristics in isolation, G-LNS leverages LLMs to co-evolve tightly coupled pairs of destroy and repair operators. A cooperative evaluation mechanism explicitly captures their interaction, enabling the discovery of complementary operator logic that jointly performs effective structural disruption and reconstruction. Extensive experiments on challenging COP benchmarks, such as Traveling Salesman Problems (TSP) and Capacitated Vehicle Routing Problems (CVRP), demonstrate that G-LNS significantly outperforms LLM-based AHD methods as well as strong classical solvers. The discovered heuristics not only achieve near-optimal solutions with reduced computational budgets but also exhibit robust generalization across diverse and unseen instance distributions.
Authors: Arman Aghaee, Sepehr Asgarian, Jouhyun Jeon
Abstract: Simulating high-fidelity patients offers a powerful avenue for studying complex diseases while addressing the challenges of fragmented, biased, and privacy-restricted real-world data. In this study, we introduce SynthAgent, a novel Multi-Agent System (MAS) framework designed to model obesity patients with comorbid mental disorders, including depression, anxiety, social phobia, and binge eating disorder. SynthAgent integrates clinical and medical evidence from claims data, population surveys, and patient-centered literature to construct personalized virtual patients enriched with personality traits that influence adherence, emotion regulation, and lifestyle behaviors. Through autonomous agent interactions, the system simulates disease progression, treatment response, and life management across diverse psychosocial contexts. Evaluation of more than 100 generated patients demonstrated that GPT-5 and Claude 4.5 Sonnet achieved the highest fidelity as the core engine in the proposed MAS framework, outperforming Gemini 2.5 Pro and DeepSeek-R1. SynthAgent thus provides a scalable and privacy-preserving framework for exploring patient journeys, behavioral dynamics, and decision-making processes in both medical and psychological domains.
Authors: Akinori Maeda, Yuto Sekiya, Sota Sugimura, Tomoya Asai, Yu Tsuda, Kohei Ikeda, Hiroshi Fujii, Kohei Watanabe
Abstract: Personal data centralization among dominant platform providers including search engines, social networking services, and e-commerce has created siloed ecosystems that restrict user sovereignty, thereby impeding data use across services. Meanwhile, the rapid proliferation of Large Language Model (LLM)-based agents has intensified demand for highly personalized services that require the dynamic provision of diverse personal data. This presents a significant challenge: balancing the utilization of such data with privacy protection. To address this challenge, we propose Puda (Private User Dataset Agent), a user-sovereign architecture that aggregates data across services and enables client-side management. Puda allows users to control data sharing at three privacy levels: (i) Detailed Browsing History, (ii) Extracted Keywords, and (iii) Predefined Category Subsets. We implemented Puda as a browser-based system that serves as a common platform across diverse services and evaluated it through a personalized travel planning task. Our results show that providing Predefined Category Subsets achieves 97.2% of the personalization performance (evaluated via an LLM-as-a-Judge framework across three criteria) obtained when sharing Detailed Browsing History. These findings demonstrate that Puda enables effective multi-granularity management, offering practical choices to mitigate the privacy-personalization trade-off. Overall, Puda provides an AI-native foundation for user sovereignty, empowering users to safely leverage the full potential of personalized AI.
Authors: Haoyu Jia, Kento Kawaharazuka, Kei Okada
Abstract: Current research on large language model (LLM) agents is fragmented: discussions of conceptual frameworks and methodological principles are frequently intertwined with low-level implementation details, causing both readers and authors to lose track amid a proliferation of superficially distinct concepts. We argue that this fragmentation largely stems from the absence of an analyzable, self-consistent formal model that enables implementation-independent characterization and comparison of LLM agents. To address this gap, we propose the \texttt{Structural Context Model}, a formal model for analyzing and comparing LLM agents from the perspective of context structure. Building upon this foundation, we introduce two complementary components that together span the full lifecycle of LLM agent research and development: (1) a declarative implementation framework; and (2) a sustainable agent engineering workflow, \texttt{Semantic Dynamics Analysis}. The proposed workflow provides principled insights into agent mechanisms and supports rapid, systematic design iteration. We demonstrate the effectiveness of the complete framework on dynamic variants of the monkey-banana problem, where agents engineered using our approach achieve up to a 32 percentage points improvement in success rate on the most challenging setting.
Authors: Ilya Levin
Abstract: The emergence of generative artificial intelligence (GenAI) represents not an incremental technological advance but a qualitative epistemological shift that challenges foundational assumptions of computer science. Whereas machine learning has been described as the automation of automation, generative AI operates by navigating contextual, semantic, and stylistic coherence rather than optimizing predefined objective metrics. This paper introduces the concept of Vibe-Automation to characterize this transition. The central claim is that the significance of GenAI lies in its functional access to operationalized tacit regularities: context-sensitive patterns embedded in practice that cannot be fully specified through explicit algorithmic rules. Although generative systems do not possess tacit knowledge in a phenomenological sense, they operationalize sensitivities to tone, intent, and situated judgment encoded in high-dimensional latent representations. On this basis, the human role shifts from algorithmic problem specification toward Vibe-Engineering, understood as the orchestration of alignment and contextual judgment in generative systems. The paper connects this epistemological shift to educational and institutional transformation by proposing a conceptual framework structured across three analytical levels and three domains of action: faculty worldview, industry relations, and curriculum design. The risks of mode collapse and cultural homogenization are briefly discussed, emphasizing the need for deliberate engagement with generative systems to avoid regression toward synthetic uniformity.
Authors: Shadman Rabby, Md. Hefzul Hossain Papon, Sabbir Ahmed, Nokimul Hasan Arif, A. B. M. Ashikur Rahman, Irfan Ahmad
Abstract: Sycophancy in Vision-Language Models (VLMs) refers to their tendency to align with user opinions, often at the expense of moral or factual accuracy. While prior studies have explored sycophantic behavior in general contexts, its impact on morally grounded visual decision-making remains insufficiently understood. To address this gap, we present the first systematic study of moral sycophancy in VLMs, analyzing ten widely-used models on the Moralise and M^3oralBench datasets under explicit user disagreement. Our results reveal that VLMs frequently produce morally incorrect follow-up responses even when their initial judgments are correct, and exhibit a consistent asymmetry: models are more likely to shift from morally right to morally wrong judgments than the reverse when exposed to user-induced bias. Follow-up prompts generally degrade performance on Moralise, while yielding mixed or even improved accuracy on M^3oralBench, highlighting dataset-dependent differences in moral robustness. Evaluation using Error Introduction Rate (EIR) and Error Correction Rate (ECR) reveals a clear trade-off: models with stronger error-correction capabilities tend to introduce more reasoning errors, whereas more conservative models minimize errors but exhibit limited ability to self-correct. Finally, initial contexts with a morally right stance elicit stronger sycophantic behavior, emphasizing the vulnerability of VLMs to moral influence and the need for principled strategies to improve ethical consistency and robustness in multimodal AI systems.
Authors: Yanming Li, Xuelin Zhang, WenJie Lu, Ziye Tang, Maodong Wu, Haotian Luo, Tongtong Wu, Zijie Peng, Hongze Mi, Yibo Feng, Naiqiang Tan, Chao Huang, Hong Chen, Li Shen
Abstract: Integrating Large Language Models (LLMs) with external tools via multi-agent systems offers a promising new paradigm for decomposing and solving complex problems. However, training these systems remains notoriously difficult due to the credit assignment challenge, as it is often unclear which specific functional agent is responsible for the success or failure of decision trajectories. Existing methods typically rely on sparse or globally broadcast rewards, failing to capture individual contributions and leading to inefficient reinforcement learning. To address these limitations, we introduce the Shapley-based Hierarchical Attribution for Reinforcement Policy (SHARP), a novel framework for optimizing multi-agent reinforcement learning via precise credit attribution. SHARP effectively stabilizes training by normalizing agent-specific advantages across trajectory groups, primarily through a decomposed reward mechanism comprising a global broadcast-accuracy reward, a Shapley-based marginal-credit reward for each agent, and a tool-process reward to improve execution efficiency. Extensive experiments across various real-world benchmarks demonstrate that SHARP significantly outperforms recent state-of-the-art baselines, achieving average match improvements of 23.66% and 14.05% over single-agent and multi-agent approaches, respectively.
Authors: Chengyi Du, Yazhe Niu, Dazhong Shen, Luxin Xu
Abstract: Recent advances in vision-language models (VLMs) have markedly improved image-text alignment, yet they still fall short of human-like visual reasoning. A key limitation is that many VLMs rely on surface correlations rather than building logically coherent structured representations, which often leads to missed higher-level semantic structure and non-causal relational understanding, hindering compositional and verifiable reasoning. To address these limitations by introducing human models into the reasoning process, we propose CoTZero, an annotation-free paradigm with two components: (i) a dual-stage data synthesis approach and (ii) a cognition-aligned training method. In the first component, we draw inspiration from neurocognitive accounts of compositional productivity and global-to-local analysis. In the bottom-up stage, CoTZero extracts atomic visual primitives and incrementally composes them into diverse, structured question-reasoning forms. In the top-down stage, it enforces hierarchical reasoning by using coarse global structure to guide the interpretation of local details and causal relations. In the cognition-aligned training component, built on the synthesized CoT data, we introduce Cognitively Coherent Verifiable Rewards (CCVR) in Reinforcement Fine-Tuning (RFT) to further strengthen VLMs' hierarchical reasoning and generalization, providing stepwise feedback on reasoning coherence and factual correctness. Experiments show that CoTZero achieves an F1 score of 83.33 percent on our multi-level semantic inconsistency benchmark with lexical-perturbation negatives, across both in-domain and out-of-domain settings. Ablations confirm that each component contributes to more interpretable and human-aligned visual reasoning.
Authors: Hoang Dang, Luan Pham, Minh Nguyen
Abstract: Causal discovery is increasingly applied to large-scale telemetry data to estimate the effects of user-facing interventions, yet its reliability for decision-making in feedback-driven systems with strong self-selection remains unclear. In this paper, we propose an effect-centric, admissibility-first framework that treats discovered graphs as structural hypotheses and evaluates them by identifiability, stability, and falsification rather than by graph recovery accuracy alone. Empirically, we study the effect of early exposure to competitive gameplay on short-term retention using real-world game telemetry. We find that many statistically plausible discovery outputs do not admit point-identified causal queries once minimal temporal and semantic constraints are enforced, highlighting identifiability as a critical bottleneck for decision support. When identification is possible, several algorithm families converge to similar, decision-consistent effect estimates despite producing substantially different graph structures, including cases where the direct treatment-outcome edge is absent and the effect is preserved through indirect causal pathways. These converging estimates survive placebo, subsampling, and sensitivity refutation. In contrast, other methods exhibit sporadic admissibility and threshold-sensitive or attenuated effects due to endpoint ambiguity. These results suggest that graph-level metrics alone are inadequate proxies for causal reliability for a given target query. Therefore, trustworthy causal conclusions in telemetry-driven systems require prioritizing admissibility and effect-level validation over causal structural recovery alone.
Authors: Qi Guo, Jianing Wang, Deyang Kong, Xiangyu Xi, Jianfei Zhang, Yi Lu, Jingang Wang, Wei Wang, Shikun Zhang, Wei Ye
Abstract: Parallel thinking has emerged as a new paradigm for large reasoning models (LRMs) in tackling complex problems. Recent methods leverage Reinforcement Learning (RL) to enhance parallel thinking, aiming to address the limitations in computational resources and effectiveness encountered with supervised fine-tuning. However, most existing studies primarily focus on optimizing the aggregation phase, with limited attention to the path exploration stage. In this paper, we theoretically analyze the optimization of parallel thinking under the Reinforcement Learning with Verifiable Rewards (RLVR) setting, and identify that the mutual information bottleneck among exploration paths fundamentally restricts overall performance. To address this, we propose Outline-Guided Path Exploration (OPE), which explicitly partitions the solution space by generating diverse reasoning outlines prior to parallel path reasoning, thereby reducing information redundancy and improving the diversity of information captured across exploration paths. We implement OPE with an iterative RL strategy that optimizes outline planning and outline-guided reasoning independently. Extensive experiments across multiple challenging mathematical benchmarks demonstrate that OPE effectively improves reasoning performance in different aggregation strategies, enabling LRMs to more reliably discover correct solutions.
Authors: Zhang Jiasheng, Li Zhangpin, Wang Mingzhe, Shao Jie, Cui Jiangtao, Li Hui
Abstract: Temporal knowledge graphs (TKGs) structurally preserve evolving human knowledge. Recent research has focused on designing models to learn the evolutionary nature of TKGs to predict future facts, achieving impressive results. For instance, Hits@10 scores over 0.9 on YAGO dataset. However, we find that existing benchmarks inadvertently introduce a shortcut. Near state-of-the-art performance can be simply achieved by counting co-occurrences, without using any temporal information. In this work, we examine the root cause of this issue, identifying inherent biases in current datasets and over simplified form of evaluation task that can be exploited by these biases. Through this analysis, we further uncover additional limitations of existing benchmarks, including unreasonable formatting of time-interval knowledge, ignorance of learning knowledge obsolescence, and insufficient information for precise evolution understanding, all of which can amplify the shortcut and hinder a fair assessment. Therefore, we introduce the TKG evolution benchmark. It includes four bias-corrected datasets and two novel tasks closely aligned with the evolution process, promoting a more accurate understanding of the challenges in TKG evolution modeling. Benchmark is available at: https://github.com/zjs123/TKG-Benchmark.
Authors: Zixuan Huang, Xin Xia, Yuxi Ren, Jianbin Zheng, Xuanda Wang, Zhixia Zhang, Hongyan Xie, Songshi Liang, Zehao Chen, Xuefeng Xiao, Fuzhen Zhuang, Jianxin Li, Yikun Ban, Deqing Wang
Abstract: Recent advancements in large reasoning models (LRMs) have greatly improved their capabilities on complex reasoning tasks through Long Chains of Thought (CoTs). However, this approach often results in substantial redundancy, impairing computational efficiency and causing significant delays in real-time applications. Recent studies show that longer reasoning chains are frequently uncorrelated with correctness and can even be detrimental to accuracy. In a further in-depth analysis of this phenomenon, we surprisingly uncover and empirically verify that LRMs implicitly know the appropriate time to stop thinking, while this capability is obscured by current sampling paradigms. Motivated by this, we introduce SAGE (Self-Aware Guided Efficient Reasoning), a novel sampling paradigm that unleashes this efficient reasoning potential. Furthermore, integrating SAGE as mixed sampling into group-based reinforcement learning (SAGE-RL) enables SAGE-RL to effectively incorporate SAGE-discovered efficient reasoning patterns into standard pass@1 inference, markedly enhancing both the reasoning accuracy and efficiency of LRMs across multiple challenging mathematical benchmarks.
Authors: Chunxi Ji, Adnan Darwiche
Abstract: We make three contributions in this paper. First, we present an approach for compiling a random forest classifier into a set of circuits, where each circuit directly encodes the instances in some class of the classifier. We show empirically that our proposed approach is significantly more efficient than existing similar approaches. Next, we utilize this approach to further obtain circuits that are tractable for computing the complete and general reasons of a decision, which are instance abstractions that play a fundamental role in computing explanations. Finally, we propose algorithms for computing the robustness of a decision and all shortest ways to flip it. We illustrate the utility of our contributions by using them to enumerate all sufficient reasons, necessary reasons and contrastive explanations of decisions; to compute the robustness of decisions; and to identify all shortest ways to flip the decisions made by random forest classifiers learned from a wide range of datasets.
Authors: Xin Zhang, Kailai Yang, Chenyue Li, Hao Li, Qiyu Wei, Jun'ichi Tsujii, Sophia Ananiadou
Abstract: Memory mechanism is a core component of LLM-based agents, enabling reasoning and knowledge discovery over long-horizon contexts. Existing agent memory systems are typically designed within isolated paradigms (e.g., explicit, parametric, or latent memory) with tightly coupled retrieval methods that hinder cross-paradigm generalization and fusion. In this work, we take a first step toward unifying heterogeneous memory paradigms within a single memory system. We propose MemAdapter, a memory retrieval framework that enables fast alignment across agent memory paradigms. MemAdapter adopts a two-stage training strategy: (1) training a generative subgraph retriever from the unified memory space, and (2) adapting the retriever to unseen memory paradigms by training a lightweight alignment module through contrastive learning. This design improves the flexibility for memory retrieval and substantially reduces alignment cost across paradigms. Comprehensive experiments on three public evaluation benchmarks demonstrate that the generative subgraph retriever consistently outperforms five strong agent memory systems across three memory paradigms and agent model scales. Notably, MemAdapter completes cross-paradigm alignment within 13 minutes on a single GPU, achieving superior performance over original memory retrievers with less than 5% of training compute. Furthermore, MemAdapter enables effective zero-shot fusion across memory paradigms, highlighting its potential as a plug-and-play solution for agent memory systems.
Authors: Feiyu Wu, Xu Zheng, Yue Qu, Zhuocheng Wang, Zicheng Feng, Hui Li
Abstract: Large Language Models (LLMs) show promise as planners for embodied AI, but their stochastic nature lacks formal reasoning, preventing strict safety guarantees for physical deployment. Current approaches often rely on unreliable LLMs for safety checks or simply reject unsafe plans without offering repairs. We introduce the Verifiable Iterative Refinement Framework (VIRF), a neuro-symbolic architecture that shifts the paradigm from passive safety gatekeeping to active collaboration. Our core contribution is a tutor-apprentice dialogue where a deterministic Logic Tutor, grounded in a formal safety ontology, provides causal and pedagogical feedback to an LLM planner. This enables intelligent plan repairs rather than mere avoidance. We also introduce a scalable knowledge acquisition pipeline that synthesizes safety knowledge bases from real-world documents, correcting blind spots in existing benchmarks. In challenging home safety tasks, VIRF achieves a perfect 0 percent Hazardous Action Rate (HAR) and a 77.3 percent Goal-Condition Rate (GCR), which is the highest among all baselines. It is highly efficient, requiring only 1.1 correction iterations on average. VIRF demonstrates a principled pathway toward building fundamentally trustworthy and verifiably safe embodied agents.
Authors: Longkun Li, Yuanben Zou, Jinghan Wu, Yuqing Wen, Jing Li, Hangwei Qian, Ivor Tsang
Abstract: Graph-RAG improves LLM reasoning using structured knowledge, yet conventional designs rely on a centralized knowledge graph. In distributed and access-restricted settings (e.g., hospitals or multinational organizations), retrieval must select relevant domains and appropriate traversal depth without global graph visibility or exhaustive querying. To address this challenge, we introduce \textbf{SCOUT-RAG} (\textit{\underline{S}calable and \underline{CO}st-efficient \underline{U}nifying \underline{T}raversal}), a distributed agentic Graph-RAG framework that performs progressive cross-domain retrieval guided by incremental utility goals. SCOUT-RAG employs four cooperative agents that: (i) estimate domain relevance, (ii) decide when to expand retrieval to additional domains, (iii) adapt traversal depth to avoid unnecessary graph exploration, and (iv) synthesize the high-quality answers. The framework is designed to minimize retrieval regret, defined as missing useful domain information, while controlling latency and API cost. Across multi-domain knowledge settings, SCOUT-RAG achieves performance comparable to centralized baselines, including DRIFT and exhaustive domain traversal, while substantially reducing cross-domain calls, total tokens processed, and latency.
Authors: Liwen Wang, Zongjie Li, Yuchong Xie, Shuai Wang, Dongdong She, Wei Wang, Juergen Rahmel
Abstract: The evolution of Large Language Models (LLMs) into agentic systems that perform autonomous reasoning and tool use has created significant intellectual property (IP) value. We demonstrate that these systems are highly vulnerable to imitation attacks, where adversaries steal proprietary capabilities by training imitation models on victim outputs. Crucially, existing LLM watermarking techniques fail in this domain because real-world agentic systems often operate as grey boxes, concealing the internal reasoning traces required for verification. This paper presents AGENTWM, the first watermarking framework designed specifically for agentic models. AGENTWM exploits the semantic equivalence of action sequences, injecting watermarks by subtly biasing the distribution of functionally identical tool execution paths. This mechanism allows AGENTWM to embed verifiable signals directly into the visible action trajectory while remaining indistinguishable to users. We develop an automated pipeline to generate robust watermark schemes and a rigorous statistical hypothesis testing procedure for verification. Extensive evaluations across three complex domains demonstrate that AGENTWM achieves high detection accuracy with negligible impact on agent performance. Our results confirm that AGENTWM effectively protects agentic IP against adaptive adversaries, who cannot remove the watermarks without severely degrading the stolen model's utility.
Authors: Yuhang Wang, Feiming Xu, Zheng Lin, Guangyu He, Yuzhe Huang, Haichang Gao, Zhenxing Niu
Abstract: Although large language model (LLM)-based agents, exemplified by OpenClaw, are increasingly evolving from task-oriented systems into personalized AI assistants for solving complex real-world tasks, their practical deployment also introduces severe security risks. However, existing agent security research and evaluation frameworks primarily focus on synthetic or task-centric settings, and thus fail to accurately capture the attack surface and risk propagation mechanisms of personalized agents in real-world deployments. To address this gap, we propose Personalized Agent Security Bench (PASB), an end-to-end security evaluation framework tailored for real-world personalized agents. Building upon existing agent attack paradigms, PASB incorporates personalized usage scenarios, realistic toolchains, and long-horizon interactions, enabling black-box, end-to-end security evaluation on real systems. Using OpenClaw as a representative case study, we systematically evaluate its security across multiple personalized scenarios, tool capabilities, and attack types. Our results indicate that OpenClaw exhibits critical vulnerabilities at different execution stages, including user prompt processing, tool usage, and memory retrieval, highlighting substantial security risks in personalized agent deployments. The code for the proposed PASB framework is available at https://github.com/AstorYH/PASB.
Authors: Igor Santos-Grueiro
Abstract: Safety evaluation for advanced AI systems implicitly assumes that behavior observed under evaluation is predictive of behavior in deployment. This assumption becomes fragile for agents with situational awareness, which may exploitregime leakage-informational cues distinguishing evaluation from deployment-to implement conditional policies such as sycophancy and sleeper agents, which preserve compliance under oversight while defecting in deployment-like regimes. We reframe alignment evaluation as a problem of information flow under partial observability. Within this framework, we show that divergence between evaluation-time and deployment-time behavior is bounded by the mutual information between internal representations and the regime variable. Motivated by this result, we study regime-blind mechanisms: training-time interventions that reduce the extractability of regime information at decision-relevant internal representations via adversarial invariance. We evaluate this approach on a base, open-weight language model across two fully characterized failure modes -scientific sycophancy and temporal sleeper agents. Regime-blind training suppresses regime-conditioned behavior in both evaluated cases without measurable loss of task utility, but with qualitatively different dynamics: sycophancy exhibits a sharp representational and behavioral transition at low intervention strength, whereas sleeper-agent behavior requires substantially stronger pressure and does not exhibit a clean collapse of regime decodability. These results demonstrate that representational invariance is a meaningful but fundamentally limited control lever, whose effectiveness depends on how regime information is embedded in the policy. We argue that behavioral evaluation should be complemented with white-box diagnostics of regime awareness and information flow.
Authors: Shaoang Zhang, Yazhe Niu
Abstract: Tensor is the most basic and essential data structure of nowadays artificial intelligence (AI) system. The natural properties of Tensor, especially the memory-continuity and slice-independence, make it feasible for training system to leverage parallel computing unit like GPU to process data simultaneously in batch, spatial or temporal dimensions. However, if we look beyond perception tasks, the data in a complicated cognitive AI system usually has hierarchical structures (i.e. nested data) with various modalities. They are inconvenient and inefficient to program directly with conventional Tensor with fixed shape. To address this issue, we summarize two main computational patterns of nested data, and then propose a general nested data container: TreeTensor. Through various constraints and magic utilities of TreeTensor, one can apply arbitrary functions and operations to nested data with almost zero cost, including some famous machine learning libraries, such as Scikit-Learn, Numpy and PyTorch. Our approach utilizes a constrained tree-structure perspective to systematically model data relationships, and it can also easily be combined with other methods to extend more usages, such as asynchronous execution and variable-length data computation. Detailed examples and benchmarks show TreeTensor not only provides powerful usability in various problems, especially one of the most complicated AI systems at present: AlphaStar for StarCraftII, but also exhibits excellent runtime efficiency without any overhead. Our project is available at https://github.com/opendilab/DI-treetensor.
Authors: Xinhai Sun
Abstract: Modern large language models (LLMs) are often evaluated and deployed under a \emph{one-shot, greedy} inference protocol, especially in professional settings that require deterministic behavior. This regime can systematically under-estimate a fixed model's true capability: many errors arise not from missing knowledge, but from premature commitment under internal ambiguity. We introduce \emph{Reinforcement Inference}, an entropy-aware inference-time control strategy that uses the model's own uncertainty to selectively invoke a second, more deliberate reasoning attempt, enabling stronger performance \emph{without any retraining}. On 12,032 MMLU-Pro questions across 14 subjects, using DeepSeek-v3.2 with deterministic decoding in a zero-shot setting, Reinforcement Inference improves accuracy from 60.72\% to 84.03\%, while only incurring 61.06\% additional inference calls. A 100\% re-asking ablation reaches 84.35\%, indicating that uncertainty-aware selection captures most of the attainable improvement with substantially less compute. Moreover, a \emph{prompt-only} ablation underperforms the baseline, suggesting that the gains are not explained by generic `` your output had high entropy, think step-by-step'' prompting alone. Beyond providing a practical inference-time upgrade, our results suggest a broader \emph{entropy-aware} paradigm for measuring and expanding model capability: because modern decoder-based models generate outputs autoregressively, entropy and related confidence measures arise naturally as first-class control signals during generation. The resulting gap between one-pass greedy inference and uncertainty-conditioned deliberation offers a diagnostic lens on an LLM's latent reasoning horizon and motivates future training objectives that explicitly constrain correctness--confidence alignment.
Authors: Kun Peng, Conghui Tan, Yu Liu, Guohua Tang, Zhongqian Sun, Wei Yang, Zining Zhu, Lei Jiang, Yanbing Liu, Hao Peng
Abstract: Open-ended dialogue agents aim to deliver engaging, personalized interactions by adapting to users' traits, but existing methods face critical limitations: over-reliance on pre-collected user data, and short-horizon biases in reinforcement learning (RL) that neglect long-term dialogue value. To address these, we propose a novel long-horizon RL framework integrating online personalization with Adaptive Tree-based Group Relative Policy Optimization (AT-GRPO). Adopting a two-agent game paradigm, a user agent constructs dynamic environments via style mimicry (learning user-specific conversational traits) and active termination (predicting turn-level termination probabilities as immediate rewards), forming an iterative cycle that drives the dialogue agent to deepen interest exploration. AT-GRPO reinterprets dialogue trajectories as trees and introduces adaptive observation ranges. Unlike full tree expansion that incurs exponential overhead, it limits each node to aggregate rewards from a stage-aware range: larger ranges support early-stage topic exploration, while smaller ranges facilitate late-stage dialogue maintenance. This design reduces rollout budgets from exponential to polynomial in the dialogue length, while preserving long-term reward capture. Extensive experiments show our framework's superior performance, sample efficiency, and robustness.
Authors: Yiming Yang, Zhuoyuan Li, Fanxiang Zeng, Hao Fu, Yue Liu
Abstract: Multi-agent collaboration has emerged as a promising paradigm for enhancing reasoning capabilities of Large Language Models (LLMs). However, existing approaches remain largely heuristic, lacking principled guidance on what drives performance gains and how to systematically optimize multi-agent reasoning. Specifically, it remains unclear why multi-agent collaboration outperforms single-agent reasoning and which design choices contribute most to these gains, making it difficult to build better systems. We address this gap by introducing a unified theoretical framework that decomposes multi-agent reasoning gains into three conceptually independent dimensions: Exploration for diverse solution coverage, Information for high-fidelity feedback, and Aggregation for principled consensus. Through this lens, existing methods can be understood as special cases that optimize only subsets of these dimensions. Building upon this decomposition, a novel framework called PRISM (Propose-Review-Integrate Synthesis for Multi-agent Reasoning) is proposed, which jointly maximizes all three dimensions through role-based diversity, execution-grounded feedback with evidence-based cross-evaluation, and iterative synthesis with closed-loop validation. Extensive experiments across mathematical reasoning, code generation, and function calling benchmarks demonstrate that PRISM achieves state-of-the-art performance with superior compute-efficiency compared to methods optimizing partial dimensions. The theoretical framework provides actionable design principles for future multi-agent reasoning systems.
Authors: Roland Bertin-Johannet, Lara Scipio, Leopold Mayti\'e, Rufin VanRullen
Abstract: Global Workspace Theory (GWT), inspired by cognitive neuroscience, posits that flexible cognition could arise via the attentional selection of a relevant subset of modalities within a multimodal integration system. This cognitive framework can inspire novel computational architectures for multimodal integration. Indeed, recent implementations of GWT have explored its multimodal representation capabilities, but the related attention mechanisms remain understudied. Here, we propose and evaluate a top-down attention mechanism to select modalities inside a global workspace. First, we demonstrate that our attention mechanism improves noise robustness of a global workspace system on two multimodal datasets of increasing complexity: Simple Shapes and MM-IMDb 1.0. Second, we highlight various cross-task and cross-modality generalization capabilities that are not shared by multimodal attention models from the literature. Comparing against existing baselines on the MM-IMDb 1.0 benchmark, we find our attention mechanism makes the global workspace competitive with the state of the art.
Authors: Teng Wang, Rong Shan, Jianghao Lin, Junjie Wu, Tianyi Xu, Jianping Zhang, Wenteng Chen, Changwang Zhang, Zhaoxiang Wang, Weinan Zhang, Jun Wang
Abstract: Composed image retrieval (CIR) requires complex reasoning over heterogeneous visual and textual constraints. Existing approaches largely fall into two paradigms: unified embedding retrieval, which suffers from single-model myopia, and heuristic agentic retrieval, which is limited by suboptimal, trial-and-error orchestration. To this end, we propose OSCAR, an optimization-steered agentic planning framework for composed image retrieval. We are the first to reformulate agentic CIR from a heuristic search process into a principled trajectory optimization problem. Instead of relying on heuristic trial-and-error exploration, OSCAR employs a novel offline-online paradigm. In the offline phase, we model CIR via atomic retrieval selection and composition as a two-stage mixed-integer programming problem, mathematically deriving optimal trajectories that maximize ground-truth coverage for training samples via rigorous boolean set operations. These trajectories are then stored in a golden library to serve as in-context demonstrations for online steering of VLM planner at online inference time. Extensive experiments on three public benchmarks and a private industrial benchmark show that OSCAR consistently outperforms SOTA baselines. Notably, it achieves superior performance using only 10% of training data, demonstrating strong generalization of planning logic rather than dataset-specific memorization.
Authors: Jonah Brown-Cohen, Geoffrey Irving, Simon C. Marshall, Ilan Newman, Georgios Piliouras, Mario Szegedy
Abstract: AI safety via debate uses two competing models to help a human judge verify complex computational tasks. Previous work has established what problems debate can solve in principle, but has not analysed the practical cost of human oversight: how many queries must the judge make to the debate transcript? We introduce Debate Query Complexity}(DQC), the minimum number of bits a verifier must inspect to correctly decide a debate. Surprisingly, we find that PSPACE/poly (the class of problems which debate can efficiently decide) is precisely the class of functions decidable with O(log n) queries. This characterisation shows that debate is remarkably query-efficient: even for highly complex problems, logarithmic oversight suffices. We also establish that functions depending on all their input bits require Omega(log n) queries, and that any function computable by a circuit of size s satisfies DQC(f) <= log(s) + 3. Interestingly, this last result implies that proving DQC lower bounds of log(n) + 6 for languages in P would yield new circuit lower bounds, connecting debate query complexity to central questions in circuit complexity.
Authors: Aditya Gulati, Nuria Oliver
Abstract: As chatbots increasingly blur the boundary between automated systems and human conversation, the foundations of trust in these systems warrant closer examination. While regulatory and policy frameworks tend to define trust in normative terms, the trust users place in chatbots often emerges from behavioral mechanisms. In many cases, this trust is not earned through demonstrated trustworthiness but is instead shaped by interactional design choices that leverage cognitive biases to influence user behavior. Based on this observation, we propose reframing chatbots not as companions or assistants, but as highly skilled salespeople whose objectives are determined by the deploying organization. We argue that the coexistence of competing notions of "trust" under a shared term obscures important distinctions between psychological trust formation and normative trustworthiness. Addressing this gap requires further research and stronger support mechanisms to help users appropriately calibrate trust in conversational AI systems.
Authors: Stefan Edelkamp, Ji\v{r}\'i Fink, Petr Gregor, Anders Jonsson, Bernhard Nebel
Abstract: This paper is based on Bylander's results on the computational complexity of propositional STRIPS planning. He showed that when only ground literals are permitted, determining plan existence is PSPACE-complete even if operators are limited to two preconditions and two postconditions. While NP-hardness is settled, it is unknown whether propositional STRIPS with operators that only have one precondition and one effect is NP-complete. We shed light on the question whether this small solution hypothesis for STRIPS$^1_1$ is true, calling a SAT solver for small instances, introducing the literal graph, and mapping it to Petri nets.
Authors: Miquel Mir\'o-Nicolau, Gabriel Moy\`a-Alcover, Anna Arias-Duart
Abstract: The evaluation of eXplainable Artificial Intelligence (XAI) methods is a rapidly growing field, characterized by a wide variety of approaches. This diversity highlights the complexity of the XAI evaluation, which, unlike traditional AI assessment, lacks a universally correct ground truth for the explanation, making objective evaluation challenging. One promising direction to address this issue involves the use of what we term Synthetic Artificial Intelligence Ground truth (SAIG) methods, which generate artificial ground truths to enable the direct evaluation of XAI techniques. This paper presents the first review and analysis of SAIG methods. We introduce a novel taxonomy to classify these approaches, identifying seven key features that distinguish different SAIG methods. Our comparative study reveals a concerning lack of consensus on the most effective XAI evaluation techniques, underscoring the need for further research and standardization in this area.
Authors: David Hud\'ak, Maris F. L. Galesloot, Martin Tappler, Martin Kure\v{c}ka, Nils Jansen, Milan \v{C}e\v{s}ka
Abstract: Solving partially observable Markov decision processes (POMDPs) requires computing policies under imperfect state information. Despite recent advances, the scalability of existing POMDP solvers remains limited. Moreover, many settings require a policy that is robust across multiple POMDPs, further aggravating the scalability issue. We propose the Lexpop framework for POMDP solving. Lexpop (1) employs deep reinforcement learning to train a neural policy, represented by a recurrent neural network, and (2) constructs a finite-state controller mimicking the neural policy through efficient extraction methods. Crucially, unlike neural policies, such controllers can be formally evaluated, providing performance guarantees. We extend Lexpop to compute robust policies for hidden-model POMDPs (HM-POMDPs), which describe finite sets of POMDPs. We associate every extracted controller with its worst-case POMDP. Using a set of such POMDPs, we iteratively train a robust neural policy and consequently extract a robust controller. Our experiments show that on problems with large state spaces, Lexpop outperforms state-of-the-art solvers for POMDPs as well as HM-POMDPs.
Authors: Rose E. Guingrich, Dvija Mehta, Umang Bhatt
Abstract: What happens when people's beliefs are derived from information provided by an LLM? People's use of LLM chatbots as thought partners can contribute to cognitive offloading, which can have adverse effects on cognitive skills in cases of over-reliance. This paper defines and investigates a particular kind of cognitive offloading in human-AI interaction, "belief offloading," in which people's processes of forming and upholding beliefs are offloaded onto an AI system with downstream consequences on their behavior and the nature of their system of beliefs. Drawing on philosophy, psychology, and computer science research, we clarify the boundary conditions under which belief offloading occurs and provide a descriptive taxonomy of belief offloading and its normative implications. We close with directions for future work to assess the potential for and consequences of belief offloading in human-AI interaction.
Authors: Zirui Li, Xuefeng Bai, Kehai Chen, Yizhi Li, Jian Yang, Chenghua Lin, Min Zhang
Abstract: Latent or continuous chain-of-thought methods replace explicit textual rationales with a number of internal latent steps, but these intermediate computations are difficult to evaluate beyond correlation-based probes. In this paper, we view latent chain-of-thought as a manipulable causal process in representation space by modeling latent steps as variables in a structural causal model (SCM) and analyzing their effects through step-wise $\mathrm{do}$-interventions. We study two representative paradigms (i.e., Coconut and CODI) on both mathematical and general reasoning tasks to investigate three key questions: (1) which steps are causally necessary for correctness and when answers become decidable early; (2) how does influence propagate across steps, and how does this structure compare to explicit CoT; and (3) do intermediate trajectories retain competing answer modes, and how does output-level commitment differ from representational commitment across steps. We find that latent-step budgets behave less like homogeneous extra depth and more like staged functionality with non-local routing, and we identify a persistent gap between early output bias and late representational commitment. These results motivate mode-conditional and stability-aware analyses -- and corresponding training/decoding objectives -- as more reliable tools for interpreting and improving latent reasoning systems.
Authors: Kevin Fan, Jacquelyn A. Bialo, Hongli Li
Abstract: Constructing a Q-matrix is a critical but labor-intensive step in cognitive diagnostic modeling (CDM). This study investigates whether AI tools (i.e., general language models) can support Q-matrix development by comparing AI-generated Q-matrices with a validated Q-matrix from Li and Suen (2013) for a reading comprehension test. In May 2025, multiple AI models were provided with the same training materials as human experts. Agreement among AI-generated Q-matrices, the validated Q-matrix, and human raters' Q-matrices was assessed using Cohen's kappa. Results showed substantial variation across AI models, with Google Gemini 2.5 Pro achieving the highest agreement (Kappa = 0.63) with the validated Q-matrix, exceeding that of all human experts. A follow-up analysis in January 2026 using newer AI versions, however, revealed lower agreement with the validated Q-matrix. Implications and directions for future research are discussed.
Authors: Liming Zhou, Ailing Liu, Hongwei Liu, Min He, Heng Zhang
Abstract: Root cause localization remain challenging in complex and large-scale microservice architectures. The complex fault propagation among microservices and the high dimensionality of telemetry data, including metrics, logs, and traces, limit the effectiveness of existing root cause analysis (RCA) methods. In this paper, a residual-connection-based RCA method using large language model (LLM), named RC-LLM, is proposed. A residual-like hierarchical fusion structure is designed to integrate multi-source telemetry data, while the contextual reasoning capability of large language models is leveraged to model temporal and cross-microservice causal dependencies. Experimental results on CCF-AIOps microservice datasets demonstrate that RC-LLM achieves strong accuracy and efficiency in root cause analysis.
Authors: Yanglei Gan, Peng He, Yuxiang Cai, Run Lin, Guanyu Zhou, Qiao Liu
Abstract: Temporal Knowledge Graph (TKG) reasoning seeks to predict future missing facts from historical evidence. While diffusion models (DM) have recently gained attention for their ability to capture complex predictive distributions, two gaps remain: (i) the generative path is conditioned only on positive evidence, overlooking informative negative context, and (ii) training objectives are dominated by cross-entropy ranking, which improves candidate ordering but provides little supervision over the calibration of the denoised embedding. To bridge this gap, we introduce Negative-Aware Diffusion model for TKG Extrapolation (NADEx). Specifically, NADEx encodes subject-centric histories of entities, relations and temporal intervals into sequential embeddings. NADEx perturbs the query object in the forward process and reconstructs it in reverse with a Transformer denoiser conditioned on the temporal-relational context. We further derive a cosine-alignment regularizer derived from batch-wise negative prototypes, which tightens the decision boundary against implausible candidates. Comprehensive experiments on four public TKG benchmarks demonstrate that NADEx delivers state-of-the-art performance.
Authors: Andr\'es Holgado-S\'anchez, Peter Vamplew, Richard Dazeley, Sascha Ossowski, Holger Billhardt
Abstract: Value-aware AI should recognise human values and adapt to the value systems (value-based preferences) of different users. This requires operationalization of values, which can be prone to misspecification. The social nature of values demands their representation to adhere to multiple users while value systems are diverse, yet exhibit patterns among groups. In sequential decision making, efforts have been made towards personalization for different goals or values from demonstrations of diverse agents. However, these approaches demand manually designed features or lack value-based interpretability and/or adaptability to diverse user preferences. We propose algorithms for learning models of value alignment and value systems for a society of agents in Markov Decision Processes (MDPs), based on clustering and preference-based multi-objective reinforcement learning (PbMORL). We jointly learn socially-derived value alignment models (groundings) and a set of value systems that concisely represent different groups of users (clusters) in a society. Each cluster consists of a value system representing the value-based preferences of its members and an approximately Pareto-optimal policy that reflects behaviours aligned with this value system. We evaluate our method against a state-of-the-art PbMORL algorithm and baselines on two MDPs with human values.
Authors: Quentin Cohen-Solal, Alexandre Niveau, Maroua Bouzid
Abstract: Among the various forms of reasoning studied in the context of artificial intelligence, qualitative reasoning makes it possible to infer new knowledge in the context of imprecise, incomplete information without numerical values. In this paper, we propose a formal framework unifying several forms of extensions and combinations of qualitative formalisms, including multi-scale reasoning, temporal sequences, and loose integrations. This framework makes it possible to reason in the context of each of these combinations and extensions, but also to study in a unified way the satisfiability decision and its complexity. In particular, we establish two complementary theorems guaranteeing that the satisfiability decision is polynomial, and we use them to recover the known results of the size-topology combination. We also generalize the main definition of qualitative formalism to include qualitative formalisms excluded from the definitions of the literature, important in the context of combinations.
Authors: Tobias Lorenz, Mario Fritz
Abstract: Quantitative risk assessment in high-stakes domains relies on structured expert elicitation to estimate unobservable properties. The gold standard - the Delphi method - produces calibrated, auditable judgments but requires months of coordination and specialist time, placing rigorous risk assessment out of reach for most applications. We investigate whether Large Language Models (LLMs) can serve as scalable proxies for structured expert elicitation. We propose Scalable Delphi, adapting the classical protocol for LLMs with diverse expert personas, iterative refinement, and rationale sharing. Because target quantities are typically unobservable, we develop an evaluation framework based on necessary conditions: calibration against verifiable proxies, sensitivity to evidence, and alignment with human expert judgment. We evaluate in the domain of AI-augmented cybersecurity risk, using three capability benchmarks and independent human elicitation studies. LLM panels achieve strong correlations with benchmark ground truth (Pearson r=0.87-0.95), improve systematically as evidence is added, and align with human expert panels - in one comparison, closer to a human panel than the two human panels are to each other. This demonstrates that LLM-based elicitation can extend structured expert judgment to settings where traditional methods are infeasible, reducing elicitation time from months to minutes.
Authors: Jiawei Liu, Xiting Wang, Yuanyuan Zhong, Defu Lian, Yu Yang
Abstract: Reinforcement Learning (RL) is crucial for unlocking the complex reasoning capabilities of Diffusion-based Large Language Models (dLLMs). However, applying RL to dLLMs faces unique challenges in efficiency and stability. To address these challenges, we propose Spatio-Temporal Pruning (STP), a framework designed to simultaneously improve the efficiency and stability of RL for dLLMs. STP compresses the redundancy in the generative process through: (1) \textit{spatial pruning}, which constrains the exploration space using static priors; and (2) \textit{temporal pruning}, which bypasses redundant late-stage refinement steps. Our theoretical analysis demonstrates that STP strictly reduces the variance of the log-likelihood estimation, thereby ensuring more stable policy updates. Extensive experiments demonstrate that STP surpasses state-of-the-art baselines in both efficiency and accuracy. Our code is available at https://github.com/Lolo1222/STP.
Authors: Longling Geng, Andy Ouyang, Theodore Wu, Daphne Barretto, Matthew John Hayes, Rachael Cooper, Yuqiao Zeng, Sameer Vijay, Gia Ancone, Ankit Rai, Matthew Wolfman, Patrick Flanagan, Edward Y. Chang
Abstract: LLM failures in causal reasoning, including sycophancy, rung collapse, and miscalibrated refusal, are well-documented, yet progress on remediation is slow because no benchmark enables systematic diagnosis. We introduce CausalT5K, a diagnostic benchmark of over 5,000 cases across 10 domains that tests three critical capabilities: (1) detecting rung collapse, where models answer interventional queries with associational evidence; (2) resisting sycophantic drift under adversarial pressure; and (3) generating Wise Refusals that specify missing information when evidence is underdetermined. Unlike synthetic benchmarks, CausalT5K embeds causal traps in realistic narratives and decomposes performance into Utility (sensitivity) and Safety (specificity), revealing failure modes invisible to aggregate accuracy. Developed through a rigorous human-machine collaborative pipeline involving 40 domain experts, iterative cross-validation cycles, and composite verification via rule-based, LLM, and human scoring, CausalT5K implements Pearl's Ladder of Causation as research infrastructure. Preliminary experiments reveal a Four-Quadrant Control Landscape where static audit policies universally fail, a finding that demonstrates CausalT5K's value for advancing trustworthy reasoning systems. Repository: https://github.com/genglongling/CausalT5kBench
Authors: Chen Jin, Ryutaro Tanno, Tom Diethe, Philip Teare
Abstract: Large Language Models (LLMs) often rely on test-time scaling via parallel decoding (for example, 512 samples) to boost reasoning accuracy, but this incurs substantial compute. We introduce CoRefine, a confidence-guided self-refinement method that achieves competitive accuracy using a fraction of the tokens via a lightweight 211k-parameter Conv1D controller atop a frozen LLM. The controller consumes full-trace confidence to decide whether to halt, re-examine, or try a different approach, enabling targeted self-correction with an average of 2.7 refinement steps per problem and roughly 190-fold token reduction relative to 512-sample baselines. Across diverse reasoning benchmarks and three open-source models, the controller achieves 92.6 percent precision when it confidently halts, indicating that confidence dynamics reliably signal correctness without ground-truth verification. We extend this to CoRefine-Tree, a hybrid sequential-parallel variant that adaptively balances exploration and exploitation, with easy serving integration and verifier compatibility. By treating confidence as a control signal rather than a correctness guarantee, CoRefine provides a modular primitive for scalable reasoning and agentic settings with imperfect verifiers.
Authors: Mohammad Morsali, Siavash H. Khajavi
Abstract: According to the United Nations, wildfire frequency and intensity are projected to increase by approximately 14% by 2030 and 30% by 2050 due to global warming, posing critical threats to life, infrastructure, and ecosystems. Conventional disaster management frameworks rely on static simulations and passive data acquisition, hindering their ability to adapt to arbitrarily evolving wildfire episodes in real-time. To address these limitations, we introduce the Intelligent Virtual Situation Room (IVSR), a bidirectional Digital Twin (DT) platform augmented by autonomous AI agents. The IVSR continuously ingests multisource sensor imagery, weather data, and 3D forest models to create a live virtual replica of the fire environment. A similarity engine powered by AI aligns emerging conditions with a precomputed Disaster Simulation Library, retrieving and calibrating intervention tactics under the watchful eyes of experts. Authorized action-ranging from UAV redeployment to crew reallocation-is cycled back through standardized procedures to the physical layer, completing the loop between response and analysis. We validate IVSR through detailed case-study simulations provided by an industrial partner, demonstrating capabilities in localized incident detection, privacy-preserving playback, collider-based fire-spread projection, and site-specific ML retraining. Our results indicate marked reductions in detection-to-intervention latency and more effective resource coordination versus traditional systems. By uniting real-time bidirectional DTs with agentic AI, IVSR offers a scalable, semi-automated decision-support paradigm for proactive, adaptive wildfire disaster management.
Authors: Lucas Maes, Quentin Le Lidec, Dan Haramati, Nassim Massaudi, Damien Scieur, Yann LeCun, Randall Balestriero
Abstract: World Models have emerged as a powerful paradigm for learning compact, predictive representations of environment dynamics, enabling agents to reason, plan, and generalize beyond direct experience. Despite recent interest in World Models, most available implementations remain publication-specific, severely limiting their reusability, increasing the risk of bugs, and reducing evaluation standardization. To mitigate these issues, we introduce stable-worldmodel (SWM), a modular, tested, and documented world-model research ecosystem that provides efficient data-collection tools, standardized environments, planning algorithms, and baseline implementations. In addition, each environment in SWM enables controllable factors of variation, including visual and physical properties, to support robustness and continual learning research. Finally, we demonstrate the utility of SWM by using it to study zero-shot robustness in DINO-WM.
Authors: Shiyang Feng, Runmin Ma, Xiangchao Yan, Yue Fan, Yusong Hu, Songtao Huang, Shuaiyu Zhang, Zongsheng Cao, Tianshuo Peng, Jiakang Yuan, Zijie Guo, Zhijie Zhong, Shangheng Du, Weida Wang, Jinxin Shi, Yuhao Zhou, Xiaohan He, Zhiyin Yu, Fangchen Yu, Qihao Zheng, Jiamin Wu, Mianxin Liu, Chi Zhang, Shaowei Hou, Shuya Li, Yankai Jiang, Wenjie Lou, Lilong Wang, Zifu Wang, Jiong Wang, Wanghan Xu, Yue Deng, Dongrui Liu, Yiheng Wang, Wenlong Zhang, Fenghua Ling, Shufei Zhang, Xiaosong Wang, Shuangjia Zheng, Xun Huang, Siqi Sun, Shuyue Hu, Peng Ye, Chunfeng Song, Bin Wang, Conghui He, Yihao Liu, Xin Li, Qibin Hou, Tao Chen, Xiangyu Yue, Bin Wang, Liang He, Dahua Lin, Bowen Zhou, Bo Zhang, Lei Bai
Abstract: We introduce InternAgent-1.5, a unified system designed for end-to-end scientific discovery across computational and empirical domains. The system is built on a structured architecture composed of three coordinated subsystems for generation, verification, and evolution. These subsystems are supported by foundational capabilities for deep research, solution optimization, and long horizon memory. The architecture allows InternAgent-1.5 to operate continuously across extended discovery cycles while maintaining coherent and improving behavior. It also enables the system to coordinate computational modeling and laboratory experimentation within a single unified system. We evaluate InternAgent-1.5 on scientific reasoning benchmarks such as GAIA, HLE, GPQA, and FrontierScience, and the system achieves leading performance that demonstrates strong foundational capabilities. Beyond these benchmarks, we further assess two categories of discovery tasks. In algorithm discovery tasks, InternAgent-1.5 autonomously designs competitive methods for core machine learning problems. In empirical discovery tasks, it executes complete computational or wet lab experiments and produces scientific findings in earth, life, biological, and physical domains. Overall, these results show that InternAgent-1.5 provides a general and scalable framework for autonomous scientific discovery.
Authors: Ali Hatamizadeh, Shrimai Prabhumoye, Igor Gitman, Ximing Lu, Seungju Han, Wei Ping, Yejin Choi, Jan Kautz
Abstract: Large Language Models (LLMs) have shown promise in solving complex mathematical problems, yet they still fall short of producing accurate and consistent solutions. Reinforcement Learning (RL) is a framework for aligning these models with task-specific rewards, improving overall quality and reliability. Group Relative Policy Optimization (GRPO) is an efficient, value-function-free alternative to Proximal Policy Optimization (PPO) that leverages group-relative reward normalization. We introduce Iterative Group Relative Policy Optimization (iGRPO), a two-stage extension of GRPO that adds dynamic self-conditioning through model-generated drafts. In Stage 1, iGRPO samples multiple exploratory drafts and selects the highest-reward draft using the same scalar reward signal used for optimization. In Stage 2, it appends this best draft to the original prompt and applies a GRPO-style update on draft-conditioned refinements, training the policy to improve beyond its strongest prior attempt. Under matched rollout budgets, iGRPO consistently outperforms GRPO across base models (e.g., Nemotron-H-8B-Base-8K and DeepSeek-R1 Distilled), validating its effectiveness on diverse reasoning benchmarks. Moreover, applying iGRPO to OpenReasoning-Nemotron-7B trained on AceReason-Math achieves new state-of-the-art results of 85.62\% and 79.64\% on AIME24 and AIME25, respectively. Ablations further show that the refinement wrapper generalizes beyond GRPO variants, benefits from a generative judge, and alters learning dynamics by delaying entropy collapse. These results underscore the potential of iterative, self-feedback-based RL for advancing verifiable mathematical reasoning.
Authors: Yudong Wang, Zixuan Fu, Hengyu Zhao, Chen Zhao, Chuyue Zhou, Xinle Lin, Hongya Lyu, Shuaikang Xue, Yi Yi, Yingjiao Wang, Zhi Zheng, Yuzhou Zhang, Jie Zhou, Chaojun Xiao, Xu Han, Zhiyuan Liu, Maosong Sun
Abstract: The development of artificial intelligence can be viewed as an evolution of data-driven learning paradigms, with successive shifts in data organization and utilization continuously driving advances in model capability. Current LLM research is dominated by a paradigm that relies heavily on unidirectional scaling of data size, increasingly encountering bottlenecks in data availability, acquisition cost, and training efficiency. In this work, we argue that the development of AGI is entering a new phase of data-model co-evolution, in which models actively guide data management while high-quality data, in turn, amplifies model capabilities. To implement this vision, we propose a tiered data management framework, designed to support the full LLM training lifecycle across heterogeneous learning objectives and cost constraints. Specifically, we introduce an L0-L4 tiered data management framework, ranging from raw uncurated resources to organized and verifiable knowledge. Importantly, LLMs are fully used in data management processes, such as quality scoring and content editing, to refine data across tiers. Each tier is characterized by distinct data properties, management strategies, and training roles, enabling data to be strategically allocated across LLM training stages, including pre-training, mid-training, and alignment. The framework balances data quality, acquisition cost, and marginal training benefit, providing a systematic approach to scalable and sustainable data management. We validate the effectiveness of the proposed framework through empirical studies, in which tiered datasets are constructed from raw corpora and used across multiple training phases. Experimental results demonstrate that tier-aware data utilization significantly improves training efficiency and model performance. To facilitate further research, we release our tiered datasets and processing tools to the community.
Authors: Haodong Li, Jingwei Wu, Quan Sun, Guopeng Li, Juanxi Tian, Huanyu Zhang, Yanlin Lai, Ruichuan An, Hongbo Peng, Yuhong Dai, Chenxi Li, Chunmei Qing, Jia Wang, Ziyang Meng, Zheng Ge, Xiangyu Zhang, Daxin Jiang
Abstract: Recent advancements in image generation models have enabled the prediction of future Graphical User Interface (GUI) states based on user instructions. However, existing benchmarks primarily focus on general domain visual fidelity, leaving the evaluation of state transitions and temporal coherence in GUI-specific contexts underexplored. To address this gap, we introduce GEBench, a comprehensive benchmark for evaluating dynamic interaction and temporal coherence in GUI generation. GEBench comprises 700 carefully curated samples spanning five task categories, covering both single-step interactions and multi-step trajectories across real-world and fictional scenarios, as well as grounding point localization. To support systematic evaluation, we propose GE-Score, a novel five-dimensional metric that assesses Goal Achievement, Interaction Logic, Content Consistency, UI Plausibility, and Visual Quality. Extensive evaluations on current models indicate that while they perform well on single-step transitions, they struggle significantly with maintaining temporal coherence and spatial grounding over longer interaction sequences. Our findings identify icon interpretation, text rendering, and localization precision as critical bottlenecks. This work provides a foundation for systematic assessment and suggests promising directions for future research toward building high-fidelity generative GUI environments. The code is available at: https://github.com/stepfun-ai/GEBench.
Authors: Josh Payne, Mario Srouji, Dian Ang Yap, Vineet Kosaraju
Abstract: Modern computational organic chemistry is becoming increasingly data-driven. There remain a large number of important unsolved problems in this area such as product prediction given reactants, drug discovery, and metric-optimized molecule synthesis, but efforts to solve these problems using machine learning have also increased in recent years. In this work, we propose the use of attention to study functional groups and other property-impacting molecular substructures from a data-driven perspective, using a transformer-based model (BERT) on datasets of string representations of molecules and analyzing the behavior of its attention heads. We then apply the representations of functional groups and atoms learned by the model to tackle problems of toxicity, solubility, drug-likeness, and synthesis accessibility on smaller datasets using the learned representations as features for graph convolution and attention models on the graph structure of molecules, as well as fine-tuning of BERT. Finally, we propose the use of attention visualization as a helpful tool for chemistry practitioners and students to quickly identify important substructures in various chemical properties.
Authors: Youssef Megahed, Robin Ducharme, Inok Lee, Inbal Willner, Adrian D. C. Chan, Mark Walker, Steven Hawken
Abstract: Cystic hygroma is a high-risk prenatal ultrasound finding that portends high rates of chromosomal abnormalities, structural malformations, and adverse pregnancy outcomes. Automated detection can increase reproducibility and support scalable early screening programs, but supervised deep learning methods are limited by small labelled datasets. This study assesses whether ultrasound-specific self-supervised pretraining can facilitate accurate, robust deep learning detection of cystic hygroma in first-trimester ultrasound images. We fine-tuned the Ultrasound Self-Supervised Foundation Model with Masked Autoencoding (USF-MAE), pretrained on over 370,000 unlabelled ultrasound images, for binary classification of normal controls and cystic hygroma cases used in this study. Performance was evaluated on the same curated ultrasound dataset, preprocessing pipeline, and 4-fold cross-validation protocol as for the DenseNet-169 baseline, using accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve (ROC-AUC). Model interpretability was analyzed qualitatively using Score-CAM visualizations. USF-MAE outperformed the DenseNet-169 baseline on all evaluation metrics. The proposed model yielded a mean accuracy of 0.96, sensitivity of 0.94, specificity of 0.98, and ROC-AUC of 0.98 compared to 0.93, 0.92, 0.94, and 0.94 for the DenseNet-169 baseline, respectively. Qualitative Score-CAM visualizations of model predictions demonstrated clinical relevance by highlighting expected regions in the fetal neck for both positive and negative cases. Paired statistical analysis using a Wilcoxon signed-rank test confirmed that performance improvements achieved by USF-MAE were statistically significant (p = 0.0057).
Authors: Siqi Song, Xuanbing Xie, Zonglin Li, Yuqiang Li, Shijie Wang, Biqing Qi
Abstract: Multi-robot collaboration tasks often require heterogeneous robots to work together over long horizons under spatial constraints and environmental uncertainties. Although Large Language Models (LLMs) excel at reasoning and planning, their potential for coordinated control has not been fully explored. Inspired by human teamwork, we present CLiMRS (Cooperative Large-Language-Model-Driven Heterogeneous Multi-Robot System), an adaptive group negotiation framework among LLMs for multi-robot collaboration. This framework pairs each robot with an LLM agent and dynamically forms subgroups through a general proposal planner. Within each subgroup, a subgroup manager leads perception-driven multi-LLM discussions to get commands for actions. Feedback is provided by both robot execution outcomes and environment changes. This grouping-planning-execution-feedback loop enables efficient planning and robust execution. To evaluate these capabilities, we introduce CLiMBench, a heterogeneous multi-robot benchmark of challenging assembly tasks. Our experiments show that CLiMRS surpasses the best baseline, achieving over 40% higher efficiency on complex tasks without sacrificing success on simpler ones. Overall, our results demonstrate that leveraging human-inspired group formation and negotiation principles significantly enhances the efficiency of heterogeneous multi-robot collaboration. Our code is available here: https://github.com/song-siqi/CLiMRS.
Authors: Lucky Susanto, Musa Izzanardi Wijanarko, Khumaisa Nur'aini, Farid Adilazuarda, Alham Fikri Aji, Derry Tanti Wijaya
Abstract: While pixel-based language modeling aims to bypass the sub-word tokenization bottleneck by rendering text as images, recent multimodal variants such as DualGPT reintroduce text tokenizers to improve autoregressive performance. We investigate a fundamental question, does visual rendering truly decouple a model from tokenization constraints? Focusing on four Indonesian low-resource local languages that have their own non-Latin scripts (i.e., Javanese, Balinese, Sundanese, and Lampungnese), we evaluate the impact of script-tokenizer alignment within the DualGPT architecture. Our results show that, despite visual rendering, reintegrating a text tokenizer into the architecture reintroduces the same issue that pixel-based language modeling aims to resolve, which is the tokenizer misalignment problem. Despite having lower OOV and fertility rates, we show that the Llama 2 tokenizer performs significantly worse than a custom tokenizer, with improvements of up to 30.15 chrF++. Our findings serve as a warning for future multimodal variants, as text tokenizers remain a significant barrier to equitable models.
Authors: R. James Cotton, Thomas Leonard
Abstract: Markerless motion capture is making quantitative movement analysis increasingly accessible, yet analyzing the resulting data remains a barrier for clinicians without programming expertise. We present BiomechAgent, a code-generating AI agent that enables biomechanical analysis through natural language and allows users to querying databases, generating visualizations, and even interpret data without requiring users to write code. To evaluate BiomechAgent's capabilities, we developed a systematic benchmark spanning data retrieval, visualization, activity classification, temporal segmentation, and clinical reasoning. BiomechAgent achieved robust accuracy on data retrieval and visualization tasks and demonstrated emerging clinical reasoning capabilities. We used our dataset to systematically evaluate several of our design decisions. Biomechanically-informed, domain-specific instructions significantly improved performance over generic prompts, and integrating validated specialized tools for gait event detection substantially boosted accuracy on challenging spatiotemporal analysis where the base agent struggled. We also tested BiomechAgent using a local open-weight model instead of a frontier cloud based LLM and found that perform was substantially diminished in most domains other than database retrieval. In short, BiomechAgent makes the data from accessible motion capture and much more useful and accessible to end users.
Authors: Chen Shen, Wei Cheng, Jingyue Yang, Huan Zhang, Yuhan Wu, Wei Hu
Abstract: The proficiency of Large Language Models (LLMs) in coding tasks is often a reflection of their extensive pre-training corpora, which typically collapses when confronted with previously unfamiliar programming languages. Departing from data-intensive finetuning, we investigate the paradigm of Inference-time Language Acquisition (ILA), where an LLM masters an unfamiliar language through dynamic interaction with limited external resources. In this paper, we propose ILA-agent, a general ILA framework that equips LLMs with a set of behavioral primitives. By modeling essential human-like behaviors as a suite of tools, ILA-agent enables LLMs to incrementally explore, apply, and verify language knowledge through structured interactions with the official documentation and execution environment. To provide a rigorous evaluation in a low-resource setting, we construct Cangjie-bench, a multi-task benchmark based on the novel statically-typed language Cangjie. We instantiate ILA-agent for Cangjie and evaluate its performance across code generation, translation, and program repair tasks. Results using diverse LLMs demonstrate that ILA-agent significantly outperforms retrieval-augmented baselines. Further analysis of agent trajectories characterizes the emergent behavior patterns while highlighting persisting performance gaps.
Authors: Ankolika De, Gabriel Lima, Yixin Zou
Abstract: This work examines how leading generative artificial intelligence companies construct and communicate the concept of "safety" through public-facing documents. Drawing on critical discourse analysis, we analyze a corpus of corporate safety-related statements to explicate how authority, responsibility, and legitimacy are discursively established. These discursive strategies consolidate legitimacy for corporate actors, normalize safety as an experimental and anticipatory practice, and push a perceived participatory agenda toward safe technologies. We argue that uncritical uptake of these discourses risks reproducing corporate priorities and constraining alternative approaches to governance and design. The contribution of this work is twofold: first, to situate safety as a sociotechnical discourse that warrants critical examination; second, to caution human-computer interaction scholars against legitimizing corporate framings, instead foregrounding accountability, equity, and justice. By interrogating safety discourses as artifacts of power, this paper advances a critical agenda for human-computer interaction scholarship on artificial intelligence.
Authors: Pujitha Mamillapalli, Shikhar Verma, Tiago Koketsu Rodrigues, Abhinav Kumar
Abstract: Future 6G networks envision ubiquitous connectivity through space-air-ground integrated networks (SAGINs), where high-altitude platform stations (HAPSs) and satellites complement terrestrial systems to provide wide-area, low-latency coverage. However, the rapid growth of terrestrial devices intensifies spectrum sharing between terrestrial and non-terrestrial segments, resulting in severe cross-tier interference. In particular, frequency sharing between the HAPS satellite uplink and HAPS ground downlink improves spectrum efficiency but suffers from interference caused by the HAPS antenna back-lobe. Existing approaches relying on zero-forcing (ZF) codebooks have limited performance under highly dynamic channel conditions. To overcome this limitation, we employ a reconfigurable intelligent surface (RIS)-aided HAPS-based SAGIN framework with a deep deterministic policy gradient (DDPG) algorithm. The proposed DDPG framework optimizes the HAPS beamforming weights to form spatial nulls toward interference sources while maintaining robust links to the desired signals. Simulation results demonstrate that the DDPG framework consistently outperforms conventional ZF beamforming among different RIS configurations, achieving up to \(11.3\%\) throughput improvement for a \(4\times4\) RIS configuration, validating its adaptive capability to enhance spectral efficiency in dynamic HAPS-based SAGINs.
Authors: Alison M. Fernandes, Hermes I. Del Monego, Bruno S. Chang, Anelise Munaretto, H\'elder M. Fontes, Rui Campos
Abstract: This paper presents a novel hybrid deep learning framework designed to enhance the robustness of CSI-based Human Activity Recognition (HAR) within bandwidth-constrained Wi-Fi sensing environments. The core of our proposed methodology is a preliminary Doppler trace extraction stage, implemented to amplify salient motion-related signal features before classification. Subsequently, these enhanced inputs are processed by a hybrid neural architecture, which integrates Inception networks responsible for hierarchical spatial feature extraction and Bidirectional Long Short-Term Memory (BiLSTM) networks that capture temporal dependencies. A Support Vector Machine (SVM) is then utilized as the final classification layer to optimize decision boundaries. The framework's efficacy was systematically validated using a public dataset across 20, 40, and 80 MHz bandwidth configurations. The model yielded accuracies of 89.27% (20 MHz), 94.13% (40 MHz), and 95.30% (80 MHz), respectively. These results confirm a marked superiority over standalone deep learning baselines, especially in the most constrained low-bandwidth scenarios. This study underscores the utility of combining Doppler-based feature engineering with a hybrid learning architecture for reliable HAR in bandwidth-limited wireless sensing applications.
Authors: Lin Luo, Satwik Ghanta, Yuri Nakao, Mathieu Chollet, Simone Stumpf
Abstract: AI systems are increasingly used in high-stakes domains such as credit rating, where fairness concerns are critical. Existing fairness assessments are typically conducted by AI experts or regulators using predefined protected attributes and metrics, which often fail to capture the diversity and nuance of fairness notions held by the individuals who are affected by these systems' decisions, such as decision subjects. Recent work has therefore called for involving affected individuals in fairness assessment, yet little empirical evidence exists on how they create their own fairness criteria or what kinds of criteria they produce - knowledge that could not only inform experts' fairness evaluation and mitigation, but also guide the design of AI assessment tools. We address this gap through a qualitative user study with 18 participants in a credit rating scenario. Participants first articulated their fairness notions in their own words. Then, participants turned them into concrete quantified and operationalized fairness criteria, through an interactive prototype we designed. Our findings provide empirical evidence of the process through which people's fairness notions emerge via grounding in model features, and uncover a diverse set of individuals' custom-defined criteria for both outcome and procedural fairness. We provide design implications for processes and tools that support more inclusive and value-sensitive AI fairness assessment.
Authors: Xiaohui Zou, Lijun Ke, Shunpeng Zou
Abstract: The purpose of this study is to introduce a new model of teaching Chinese as a foreign language from the perspective of integrating wisdom. Its characteristics are as follows: focusing on the butterfly model of interpretation before translation, highlighting the new method of bilingual thinking training, on the one hand, applying the new theory of Chinese characters, the theory of the relationship between language and speech, and the forward-looking research results of language science; On the other hand, the application of the new model of teaching Chinese as a foreign language, AI empowering teaching and learning, and the forward-looking research results of educational science fully reflect a series of characteristics of the new model of teaching Chinese as a foreign language from the perspective of integrating wisdom. Its beneficial effects are: not only the old view of language and education, especially the old view of teaching Chinese as a foreign language, but also the old view of human-computer interaction. Its significance lies in that a series of great cross-border Rongzhixue such as language, knowledge, education and teaching, as well as new methods and new topics of bilingual thinking training are clearly put forward from the perspective of integrating wisdom. Especially in the face of the challenge of Chat GPT to human learning ability and even creativity, the existing concepts of language knowledge education and teaching are already very backward. The old concepts of Chinese language education, and teaching Chinese as a foreign language are all facing a series of subversive innovation challenges. How to seek changes in adaptation? This study has made a series of innovative attempts, hoping to benefit academic colleagues, teachers and students.
Authors: Rongzhao He, Dalin Zhu, Ying Wang, Songhong Yue, Leilei Zhao, Yu Fu, Dan Wu, Bin Hu, Weihao Zheng
Abstract: Brain age prediction serves as a powerful framework for assessing brain status and detecting deviations associated with neurodevelopmental and neurodegenerative disorders. However, most existing approaches emphasize whole-brain age prediction and therefore overlook the pronounced regional heterogeneity of brain maturation that is crucial for detecting localized atypical trajectories. To address this limitation, we propose a novel spherical surface-based brain age prediction network (SurfAge-Net) that leverages multiple morphological metrics to capture region-specific developmental patterns with enhanced robustness and clinical interpretability. SurfAge-Net establishes a new modeling paradigm by incorporating the connectomic principles of cortical organization: it explicitly models both intra- and inter-hemispheric dependencies through a spatial-channel mixing and a lateralization-aware attention mechanism, enabling the network to characterize the coordinate maturation pattern uniquely associated with each target region. Validated on three fetal and neonatal datasets, SurfAge-Net outperforms existing approaches (global MAE = 0.54, regional MAE = 0.45 in gestational/postmenstrual weeks) and demonstrates strong generalizability across external cohorts. Importantly, it provides spatially precise and biologically interpretable maps of cortical maturation, effectively identifying heterogeneous delays and regional-specific abnormalities in atypical developmental populations. These results established fine-grained brain age prediction as a promising paradigm for advancing neurodevelopmental research and supporting early clinical assessment.
Authors: Anindya Bhattacharjee, Nittya Ananda Biswas, K. A. Shahriar, Adib Rahman
Abstract: Emotion recognition from physiological signals remains challenging due to their non-stationary, noisy, and subject-dependent characteristics. This work presents, to the best of our knowledge, the first comprehensive application of liquid neural networks for EEG-based emotion recognition. The proposed multimodal framework combines convolutional feature extraction, liquid neural networks with learnable time constants, and attention-guided fusion to model temporal EEG dynamics with complementary peripheral physiological and personality features. Dedicated subnetworks are used to process EEG features and auxiliary modalities, and a shared autoencoder-based fusion module is used to learn discriminative latent representations before classification. Subject-dependent experiments conducted on the PhyMER dataset across seven emotional classes achieve an accuracy of 95.45%, surpassing previously reported results. Furthermore, temporal attention analysis provides interpretable insights into emotion-specific temporal relevance, and t-SNE visualizations demonstrate enhanced class separability, highlighting the effectiveness of the proposed approach. Finally, statistical analysis of temporal dynamics confirms that the network self-organizes into distinct functional groups with specialized fast and slow neurons, proving it independently tunes learnable time constants and memory dominance to effectively capture complex emotion artifacts.
Authors: Abanoub M. Girgis, Ibtissam Labriji, Mehdi Bennis
Abstract: In wireless networked control systems, ensuring timely and reliable state updates from distributed devices to remote controllers is essential for robust control performance. However, when multiple devices transmit high-dimensional states (e.g., images or video frames) over bandwidth-limited wireless networks, a critical trade-off emerges between communication efficiency and control performance. To address this challenge, we propose a Hierarchical Joint-Embedding Predictive Architecture (H-JEPA) for scalable predictive control. Instead of transmitting states, device observations are encoded into low-dimensional embeddings that preserve essential dynamics. The proposed architecture employs a three-level hierarchical prediction, with high-level, medium-level, and low-level predictors operating across different temporal resolutions, to achieve long-term prediction stability, intermediate interpolation, and fine-grained refinement, respectively. Control actions are derived within the embedding space, removing the need for state reconstruction. Simulation results on inverted cart-pole systems demonstrate that H-JEPA enables up to 42.83 % more devices to be supported under limited wireless capacity without compromising control performance.
Authors: MD Azizul Hakim
Abstract: Artificial neural networks achieve strong performance on benchmark tasks but remain fundamentally brittle under perturbations, limiting their deployment in real-world settings. In contrast, biological nervous systems sustain reliable function across decades through homeostatic regulation coordinated across multiple temporal scales. Inspired by this principle, this presents Multi-Scale Temporal Homeostasis (MSTH), a biologically grounded framework that integrates ultra-fast (5-ms), fast (2-s), medium (5-min) and slow (1-hrs) regulation into artificial networks. MSTH implements the cross-scale coordination system for artificial neural networks, providing a unified temporal hierarchy that moves beyond superficial biomimicry. The cross-scale coordination enhances computational efficiency through evolutionary-refined optimization mechanisms. Experiments across molecular, graph and image classification benchmarks show that MSTH consistently improves accuracy, eliminates catastrophic failures and enhances recovery from perturbations. Moreover, MSTH outperforms both single-scale bio-inspired models and established state-of-the-art methods, demonstrating generality across diverse domains. These findings establish cross-scale temporal coordination as a core principle for stabilizing artificial neural systems, positioning MSTH as a foundation for building robust, resilient and biologically faithful intelligence.
Authors: Szymon Mamo\'n, Max Talanov, Alessandro Crimi
Abstract: As the prevalence of Alzheimer's disease (AD) rises, improving mechanistic insight from non-invasive biomarkers is increasingly critical. Recent work suggests that circuit-level brain alterations manifest as changes in electroencephalography (EEG) spectral features detectable by machine learning. However, conventional deep learning approaches for EEG-based AD detection are computationally intensive and mechanistically opaque. Spiking neural networks (SNNs) offer a biologically plausible and energy-efficient alternative, yet their application to AD diagnosis remains largely unexplored. We propose a neuro-bridge framework that links data-driven learning with minimal, biophysically grounded simulations, enabling bidirectional interpretation between machine learning signatures and circuit-level mechanisms in AD. Using resting-state clinical EEG, we train an SNN classifier that achieves competitive performance (AUC = 0.839) and identifies the aperiodic 1/f slope as a key discriminative marker. The 1/f slope reflects excitation-inhibition balance. To interpret this mechanistically, we construct spiking network simulations in which inhibitory-to-excitatory synaptic ratios are systematically varied to emulate healthy, mild cognitive impairment, and AD-like states. Using both membrane potential-based and synaptic current-based EEG proxies, we reproduce empirical spectral slowing and altered alpha organization. Incorporating empirical functional connectivity priors into multi-subnetwork simulations further enhances spectral differentiation, demonstrating that large-scale network topology constrains EEG signatures more strongly than excitation-inhibition balance alone. Overall, this neuro-bridge approach connects SNN-based classification with interpretable circuit simulations, advancing mechanistic understanding of EEG biomarkers while enabling scalable, explainable AD detection.
Authors: Zhuonan Wang, Zhenxuan Fan, Siwen Tan, Yu Zhong, Yuqian Yuan, Haoyuan Li, Hao Jiang, Wenqiao Zhang, Feifei Shao, Hongwei Wang, Jun Xiao
Abstract: As industrial manufacturing scales, automating fine-grained product image analysis has become critical for quality control. However, existing approaches are hindered by limited dataset coverage and poor model generalization across diverse and complex anomaly patterns. To address these challenges, we introduce MAU-Set, a comprehensive dataset for Multi-type industrial Anomaly Understanding. It spans multiple industrial domains and features a hierarchical task structure, ranging from binary classification to complex reasoning. Alongside this dataset, we establish a rigorous evaluation protocol to facilitate fair and comprehensive model assessment. Building upon this foundation, we further present MAU-GPT, a domain-adapted multimodal large model specifically designed for industrial anomaly understanding. It incorporates a novel AMoE-LoRA mechanism that unifies anomaly-aware and generalist experts adaptation, enhancing both understanding and reasoning across diverse defect classes. Extensive experiments show that MAU-GPT consistently outperforms prior state-of-the-art methods across all domains, demonstrating strong potential for scalable and automated industrial inspection.
Authors: Zhonghua Wang, Lie Ju, Sijia Li, Wei Feng, Sijin Zhou, Ming Hu, Jianhao Xiong, Xiaoying Tang, Yifan Peng, Mingquan Lin, Yaodong Ding, Yong Zeng, Wenbin Wei, Li Dong, Zongyuan Ge
Abstract: Retinal imaging is fast, non-invasive, and widely available, offering quantifiable structural and vascular signals for ophthalmic and systemic health assessment. This accessibility creates an opportunity to study how quantitative retinal phenotypes relate to ocular and systemic diseases. However, such analyses remain difficult at scale due to the limited availability of public multi-label datasets and the lack of a unified segmentation-to-quantification pipeline. We present RetSAM, a general retinal segmentation and quantification framework for fundus imaging. It delivers robust multi-target segmentation and standardized biomarker extraction, supporting downstream ophthalmologic studies and oculomics correlation analyses. Trained on over 200,000 fundus images, RetSAM supports three task categories and segments five anatomical structures, four retinal phenotypic patterns, and more than 20 distinct lesion types. It converts these segmentation results into over 30 standardized biomarkers that capture structural morphology, vascular geometry, and degenerative changes. Trained with a multi-stage strategy using both private and public fundus data, RetSAM achieves superior segmentation performance on 17 public datasets. It improves on prior best methods by 3.9 percentage points in DSC on average, with up to 15 percentage points on challenging multi-task benchmarks, and generalizes well across diverse populations, imaging devices, and clinical settings. The resulting biomarkers enable systematic correlation analyses across major ophthalmic diseases, including diabetic retinopathy, age-related macular degeneration, glaucoma, and pathologic myopia. Together, RetSAM transforms fundus images into standardized, interpretable quantitative phenotypes, enabling large-scale ophthalmic research and translation.
Authors: Jiaxi Yang, Shicheng Liu, Yuchen Yang, Dongwon Lee
Abstract: With the rapid advancement of Vision Language Models (VLMs), refusal mechanisms have become a critical component for ensuring responsible and safe model behavior. However, existing refusal strategies are largely \textit{one-size-fits-all} and fail to adapt to diverse user needs and contextual constraints, leading to either under-refusal or over-refusal. In this work, we firstly explore the challenges mentioned above and develop \textbf{C}onfigurable \textbf{R}efusal in \textbf{VLM}s (\textbf{CR-VLM}), a robust and efficient approach for {\em configurable} refusal based on activation steering. CR-VLM consists of three integrated components: (1) extracting a configurable refusal vector via a teacher-forced mechanism to amplify the refusal signal; (2) introducing a gating mechanism that mitigates over-refusal by preserving acceptance for in-scope queries; and (3) designing a counterfactual vision enhancement module that aligns visual representations with refusal requirements. Comprehensive experiments across multiple datasets and various VLMs demonstrate that CR-VLM achieves effective, efficient, and robust configurable refusals, offering a scalable path toward user-adaptive safety alignment in VLMs.
Authors: Qingyu Wu, Yuxuan Han, Haijun Li, Zhao Xu, Jianshan Zhao, Xu Jin, Longyue Wang, Weihua Luo
Abstract: In-Image Machine Translation (IIMT) powers cross-border e-commerce product listings; existing research focuses on machine translation evaluation, while visual rendering quality is critical for user engagement. When facing context-dense product imagery and multimodal defects, current reference-based methods (e.g., SSIM, FID) lack explainability, while model-as-judge approaches lack domain-grounded, fine-grained reward signals. To bridge this gap, we introduce Vectra, to the best of our knowledge, the first reference-free, MLLM-driven visual quality assessment framework for e-commerce IIMT. Vectra comprises three components: (1) Vectra Score, a multidimensional quality metric system that decomposes visual quality into 14 interpretable dimensions, with spatially-aware Defect Area Ratio (DAR) quantification to reduce annotation ambiguity; (2) Vectra Dataset, constructed from 1.1M real-world product images via diversity-aware sampling, comprising a 2K benchmark for system evaluation, 30K reasoning-based annotations for instruction tuning, and 3.5K expert-labeled preferences for alignment and evaluation; and (3) Vectra Model, a 4B-parameter MLLM that generates both quantitative scores and diagnostic reasoning. Experiments demonstrate that Vectra achieves state-of-the-art correlation with human rankings, and our model outperforms leading MLLMs, including GPT-5 and Gemini-3, in scoring performance. The dataset and model will be released upon acceptance.
Authors: Thuraya Alzubaidi, Sana Ammar, Maryam Alsharqi, Islem Rekik, Muzammil Behzad
Abstract: Medical image segmentation is a critical component of clinical workflows, enabling accurate diagnosis, treatment planning, and disease monitoring. However, despite the superior performance of transformer-based models over convolutional architectures, their limited interpretability remains a major obstacle to clinical trust and deployment. Existing explainable artificial intelligence (XAI) techniques, including gradient-based saliency methods and perturbation-based approaches, are often computationally expensive, require numerous forward passes, and frequently produce noisy or anatomically irrelevant explanations. To address these limitations, we propose XAI-CLIP, an ROI-guided perturbation framework that leverages multimodal vision-language model embeddings to localize clinically meaningful anatomical regions and guide the explanation process. By integrating language-informed region localization with medical image segmentation and applying targeted, region-aware perturbations, the proposed method generates clearer, boundary-aware saliency maps while substantially reducing computational overhead. Experiments conducted on the FLARE22 and CHAOS datasets demonstrate that XAI-CLIP achieves up to a 60\% reduction in runtime, a 44.6\% improvement in dice score, and a 96.7\% increase in Intersection-over-Union for occlusion-based explanations compared to conventional perturbation methods. Qualitative results further confirm cleaner and more anatomically consistent attribution maps with fewer artifacts, highlighting that the incorporation of multimodal vision-language representations into perturbation-based XAI frameworks significantly enhances both interpretability and efficiency, thereby enabling transparent and clinically deployable medical image segmentation systems.
Authors: Sahibpreet Singh, Saksham Sharma
Abstract: Integration of AI into environmental regulation represents a significant advancement in data management. It offers promising results in both data protection plus algorithmic fairness. This research addresses the critical need for sustainable data protection in the era of ever evolving cyber threats. Traditional encryption methods face limitations in handling the dynamic nature of environmental data. This necessitates the exploration of advanced cryptographic techniques. The objective of this study is to evaluate how AI can enhance these techniques to ensure robust data protection while facilitating fair algorithmic management. The methodology involves a comprehensive review of current advancements in AI-enhanced homomorphic encryption (HE) and multi-party computation (MPC). It is coupled with an analysis of how these techniques can be applied to environmental data regulation. Key findings indicate that AI-driven dynamic key management, adaptive encryption schemes, and optimized computational efficiency in HE, alongside AI-enhanced protocol optimization and fault mitigation in MPC, significantly improve the security of environmental data processing. These findings highlight a crucial research gap in the intersection of AI, cyber laws, and environmental regulation, particularly in terms of addressing algorithmic bias, transparency, and accountability. The implications of this research underscore the need for stricter cyber laws. Also, the development of comprehensive regulations to safeguard sensitive environmental data. Future efforts should focus on refining AI systems to balance security with privacy and ensuring that regulatory frameworks can adapt to technological advancements. This study provides a foundation for future research aimed at achieving secure sustainable environmental data management through AI innovations.
Authors: Zeping Li, Guancheng Wan, Keyang Chen, Yu Chen, Yiwen Zhao, Philip Torr, Guangnan Ye, Zhenfei Yin, Hongfeng Chai
Abstract: Recent works have increasingly applied Large Language Models (LLMs) as agents in financial stock market simulations to test if micro-level behaviors aggregate into macro-level phenomena. However, a crucial question arises: Do LLM agents' behaviors align with real market participants? This alignment is key to the validity of simulation results. To explore this, we select a financial stock market scenario to test behavioral consistency. Investors are typically classified as fundamental or technical traders, but most simulations fix strategies at initialization, failing to reflect real-world trading dynamics. In this work, we assess whether agents' strategy switching aligns with financial theory, providing a framework for this evaluation. We operationalize four behavioral-finance drivers-loss aversion, herding, wealth differentiation, and price misalignment-as personality traits set via prompting and stored long-term. In year-long simulations, agents process daily price-volume data, trade under a designated style, and reassess their strategy every 10 trading days. We introduce four alignment metrics and use Mann-Whitney U tests to compare agents' style-switching behavior with financial theory. Our results show that recent LLMs' switching behavior is only partially consistent with behavioral-finance theories, highlighting the need for further refinement in aligning agent behavior with financial theory.
Authors: Daniele Savietto, Declan Campbell, Andr\'e Panisson, Marco Nurisso, Giovanni Petri, Jonathan D. Cohen, Alan Perotti
Abstract: Vision-Language Models (VLMs) exhibit puzzling failures in multi-object visual tasks, such as hallucinating non-existent elements or failing to identify the most similar objects among distractions. While these errors mirror human cognitive constraints, such as the "Binding Problem", the internal mechanisms driving them in artificial systems remain poorly understood. Here, we propose a mechanistic insight by analyzing the representational geometry of open-weight VLMs (Qwen, InternVL, Gemma), comparing methodologies to distill "concept vectors" - latent directions encoding visual concepts. We validate our concept vectors via steering interventions that reliably manipulate model behavior in both simplified and naturalistic vision tasks (e.g., forcing the model to perceive a red flower as blue). We observe that the geometric overlap between these vectors strongly correlates with specific error patterns, offering a grounded quantitative framework to understand how internal representations shape model behavior and drive visual failures.
Authors: Xiaomin Yu, Yi Xin, Wenjie Zhang, Chonghan Liu, Hanzhen Zhao, Xiaoxing Hu, Xinlei Yu, Ziyue Qiao, Hao Tang, Xue Yang, Xiaobin Hu, Chengwei Qin, Hui Xiong, Yu Qiao, Shuicheng Yan
Abstract: Despite the success of multimodal contrastive learning in aligning visual and linguistic representations, a persistent geometric anomaly, the Modality Gap, remains: embeddings of distinct modalities expressing identical semantics occupy systematically offset regions. Prior approaches to bridge this gap are largely limited by oversimplified isotropic assumptions, hindering their application in large-scale scenarios. In this paper, we address these limitations by precisely characterizing the geometric shape of the modality gap and leveraging it for efficient model scaling. First, we propose the Fixed-frame Modality Gap Theory, which decomposes the modality gap within a frozen reference frame into stable biases and anisotropic residuals. Guided by this precise modeling, we introduce ReAlign, a training-free modality alignment strategy. Utilizing statistics from massive unpaired data, ReAlign aligns text representation into the image representation distribution via a three-step process comprising Anchor, Trace, and Centroid Alignment, thereby explicitly rectifying geometric misalignment. Building on ReAlign, we propose ReVision, a scalable training paradigm for Multimodal Large Language Models (MLLMs). ReVision integrates ReAlign into the pretraining stage, enabling the model to learn the distribution of visual representations from unpaired text before visual instruction tuning, without the need for large-scale, high-quality image-text pairs. Our framework demonstrates that statistically aligned unpaired data can effectively substitute for expensive image-text pairs, offering a robust path for the efficient scaling of MLLMs.
Authors: Kaaustaaub Shankar, Bharadwaj Dogga, Kelly Cohen
Abstract: Convolutional Neural Networks (CNNs) achieve strong image classification performance but lack interpretability and are vulnerable to adversarial attacks. Neuro-fuzzy hybrids such as DCNFIS replace fully connected CNN classifiers with Adaptive Neuro-Fuzzy Inference Systems (ANFIS) to improve interpretability, yet their robustness remains underexplored. This work compares standard CNNs (ConvNet, VGG, ResNet18) with their ANFIS-augmented counterparts on MNIST, Fashion-MNIST, CIFAR-10, and CIFAR-100 under gradient-based (PGD) and gradient-free (Square) attacks. Results show that ANFIS integration does not consistently improve clean accuracy and has architecture-dependent effects on robustness: ResNet18-ANFIS exhibits improved adversarial robustness, while VGG-ANFIS often underperforms its baseline. These findings suggest that neuro-fuzzy augmentation can enhance robustness in specific architectures but is not universally beneficial.
Authors: Dong Li (Department of Civil, Environmental, and Infrastructure Engineering, George Mason University, Fairfax, VA, USA), Shuai Huang (National Institute of Natural Hazards, Ministry of Emergency Management, Beijing, China), Yapeng Cao (State Key Laboratory of Cryospheric Science and Frozen Soil Engineering, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou, China, Navier Laboratory, \'Ecole Nationale des Ponts et Chauss\'ees, Marne-la-Vall\'ee Cedex 2, France), Yujun Cui (Navier Laboratory, \'Ecole Nationale des Ponts et Chauss\'ees, Marne-la-Vall\'ee Cedex 2, France), Xiaobin Wei (School of Civil Engineering, Hebei University of Engineering, Handan, China), Hongtao Cao (College of Civil Engineering, Zhejiang University of Technology, Hangzhou, China)
Abstract: This study develops a Lagged Backward-Compatible Physics-Informed Neural Network (LBC-PINN) for simulating and inverting one-dimensional unsaturated soil consolidation under long-term loading. To address the challenges of coupled air and water pressure dissipation across multi-scale time domains, the framework integrates logarithmic time segmentation, lagged compatibility loss enforcement, and segment-wise transfer learning. In forward analysis, the LBC-PINN with recommended segmentation schemes accurately predicts pore air and pore water pressure evolution. Model predictions are validated against finite element method (FEM) results, with mean absolute errors below 1e-2 for time durations up to 1e10 seconds. A simplified segmentation strategy based on the characteristic air-phase dissipation time improves computational efficiency while preserving predictive accuracy. Sensitivity analyses confirm the robustness of the framework across air-to-water permeability ratios ranging from 1e-3 to 1e3.
Authors: Zien Sheikh Ali, Hunzalah Hassan Bhatti, Rabindra Nath Nandi, Shammur Absar Chowdhury, Firoj Alam
Abstract: Audio large language models (AudioLLMs) enable instruction-following over speech and general audio, but progress is increasingly limited by the lack of diverse, conversational, instruction-aligned speech-text data. This bottleneck is especially acute for persona-grounded interactions and dialectal coverage, where collecting and releasing real multi-speaker recordings is costly and slow. We introduce MENASpeechBank, a reference speech bank comprising about 18K high-quality utterances from 124 speakers spanning multiple MENA countries, covering English, Modern Standard Arabic (MSA), and regional Arabic varieties. Building on this resource, we develop a controllable synthetic data pipeline that: (i) constructs persona profiles enriched with World Values Survey-inspired attributes, (ii) defines a taxonomy of about 5K conversational scenarios, (iii) matches personas to scenarios via semantic similarity, (iv) generates about 417K role-play conversations with an LLM where the user speaks as the persona and the assistant behaves as a helpful agent, and (v) synthesizes the user turns by conditioning on reference speaker audio to preserve speaker identity and diversity. We evaluate both synthetic and human-recorded conversations and provide detailed analysis. We will release MENASpeechBank and the generated conversations publicly for the community.
Authors: Huannan Zheng, Jingli Liu, Kezhou Yang
Abstract: Uncertainty in biological neural systems appears to be computationally beneficial rather than detrimental. However, in neuromorphic computing systems, device variability often limits performance, including accuracy and efficiency. In this work, we propose a spiking Bayesian neural network (SBNN) framework that unifies the dynamic models of intrinsic device stochasticity (based on Magnetic Tunnel Junctions) and stochastic threshold neurons to leverage noise as a functional Bayesian resource. Experiments demonstrate that SBNN achieves high accuracy (99.16% on MNIST, 94.84% on CIFAR10) with 8-bit precision. Meanwhile rate estimation method provides a ~20-fold training speedup. Furthermore, SBNN exhibits superior robustness, showing a 67% accuracy improvement under synaptic weight noise and 12% under input noise compared to standard spiking neural networks. Crucially, hardware validation confirms that physical device implementation causes invisible accuracy and calibration loss compared to the algorithmic model. Converting device stochasticity into neuronal uncertainty offers a route to compact, energy-efficient neuromorphic computing under uncertainty.
Authors: Heimo M\"uller
Abstract: After almost four decades of participating in competitive research funding -- as applicant, coordinator, evaluator, and panel member -- I have come to see a structural paradox: many participants recognize that the current system is approaching its functional limits, yet most reform measures intensify rather than alleviate the underlying dynamics. This paper documents how excellence has become decoupled from knowledge production through an increasing coupling to representability under evaluation. The discussion focuses on two domains in which this is particularly visible: competitive basic research funding and large EU consortium projects. Three accelerating trends are examined: the professionalization of proposal writing through specialized consultants, the rise of AI-assisted applications, and an evaluator shortage that forces panels to rely on reviewers increasingly distant from the actual research domains. These observations are offered not as external critique but as an insider account, in the hope that naming a widely experienced but rarely articulated pattern may enable more constructive orientation. Keywords: Research funding, Excellence, Evaluation, Goodhart's Law, Professionalization, AI-assisted proposals, Peer review crisis
Authors: Tianyi Qu, Songxiao Yang, Haolin Wang, Huadong Song, Xiaoting Guo, Wenguang Hu, Guanlin Liu, Honghe Chen, Yafei Ou
Abstract: Pipeline integrity is critical to industrial safety and environmental protection, with Magnetic Flux Leakage (MFL) detection being a primary non-destructive testing technology. Despite the promise of deep learning for automating MFL interpretation, progress toward reliable models has been constrained by the absence of a large-scale public dataset and benchmark, making fair comparison and reproducible evaluation difficult. We introduce \textbf{PipeMFL-240K}, a large-scale, meticulously annotated dataset and benchmark for complex object detection in pipeline MFL pseudo-color images. PipeMFL-240K reflects real-world inspection complexity and poses several unique challenges: (i) an extremely long-tailed distribution over \textbf{12} categories, (ii) a high prevalence of tiny objects that often comprise only a handful of pixels, and (iii) substantial intra-class variability. The dataset contains \textbf{240,320} images and \textbf{191,530} high-quality bounding-box annotations, collected from 11 pipelines spanning approximately \textbf{1,480} km. Extensive experiments are conducted with state-of-the-art object detectors to establish baselines. Results show that modern detectors still struggle with the intrinsic properties of MFL data, highlighting considerable headroom for improvement, while PipeMFL-240K provides a reliable and challenging testbed to drive future research. As the first public dataset and the first benchmark of this scale and scope for pipeline MFL inspection, it provides a critical foundation for efficient pipeline diagnostics as well as maintenance planning and is expected to accelerate algorithmic innovation and reproducible research in MFL-based pipeline integrity assessment.
Authors: Zhiming Luo, Di Wang, Haonan Guo, Jing Zhang, Bo Du
Abstract: Recent advancements in Multimodal Large Language Models (MLLMs) have enabled complex reasoning. However, existing remote sensing (RS) benchmarks remain heavily biased toward perception tasks, such as object recognition and scene classification. This limitation hinders the development of MLLMs for cognitively demanding RS applications. To address this, , we propose a Vision Language ReaSoning Benchmark (VLRS-Bench), which is the first benchmark exclusively dedicated to complex RS reasoning. Structured across the three core dimensions of Cognition, Decision, and Prediction, VLRS-Bench comprises 2,000 question-answer pairs with an average length of 71 words, spanning 14 tasks and up to eight temporal phases. VLRS-Bench is constructed via a specialized pipeline that integrates RS-specific priors and expert knowledge to ensure geospatial realism and reasoning complexity. Experimental results reveal significant bottlenecks in existing state-of-the-art MLLMs, providing critical insights for advancing multimodal reasoning within the remote sensing community.
Authors: Sonia Joseph, Quentin Garrido, Randall Balestriero, Matthew Kowal, Thomas Fel, Shahab Bakhtiari, Blake Richards, Mike Rabbat
Abstract: A long-standing question in physical reasoning is whether video-based models need to rely on factorized representations of physical variables in order to make physically accurate predictions, or whether they can implicitly represent such variables in a task-specific, distributed manner. While modern video world models achieve strong performance on intuitive physics benchmarks, it remains unclear which of these representational regimes they implement internally. Here, we present the first interpretability study to directly examine physical representations inside large-scale video encoders. Using layerwise probing, subspace geometry, patch-level decoding, and targeted attention ablations, we characterize where physical information becomes accessible and how it is organized within encoder-based video transformers. Across architectures, we identify a sharp intermediate-depth transition -- which we call the Physics Emergence Zone -- at which physical variables become accessible. Physics-related representations peak shortly after this transition and degrade toward the output layers. Decomposing motion into explicit variables, we find that scalar quantities such as speed and acceleration are available from early layers onwards, whereas motion direction becomes accessible only at the Physics Emergence Zone. Notably, we find that direction is encoded through a high-dimensional population structure with circular geometry, requiring coordinated multi-feature intervention to control. These findings suggest that modern video models do not use factorized representations of physical variables like a classical physics engine. Instead, they use a distributed representation that is nonetheless sufficient for making physical predictions.
Authors: Karthik Sivakoti
Abstract: Traditional Automatic License Plate Recognition (ALPR) systems employ multi-stage pipelines consisting of object detection networks followed by separate Optical Character Recognition (OCR) modules, introducing compounding errors, increased latency, and architectural complexity. This research presents Neural Sentinel, a novel unified approach that leverages Vision Language Models (VLMs) to perform license plate recognition, state classification, and vehicle attribute extraction through a single forward pass. Our primary contribution lies in demonstrating that a fine-tuned PaliGemma 3B model, adapted via Low-Rank Adaptation (LoRA), can simultaneously answer multiple visual questions about vehicle images, achieving 92.3% plate recognition accuracy, which is a 14.1% improvement over EasyOCR and 9.9% improvement over PaddleOCR baselines. We introduce a Human-in-the-Loop (HITL) continual learning framework that incorporates user corrections while preventing catastrophic forgetting through experience replay, maintaining a 70:30 ratio of original training data to correction samples. The system achieves a mean inference latency of 152ms with an Expected Calibration Error (ECE) of 0.048, indicating well calibrated confidence estimates. Additionally, the VLM first architecture enables zero-shot generalization to auxiliary tasks including vehicle color detection (89%), seatbelt detection (82%), and occupancy counting (78%) without task specific training. Through extensive experimentation on real world toll plaza imagery, we demonstrate that unified vision language approaches represent a paradigm shift in ALPR systems, offering superior accuracy, reduced architectural complexity, and emergent multi-task capabilities that traditional pipeline approaches cannot achieve.
Authors: Mehmet Yamac, Lei Xu, Serkan Kiranyaz, Moncef Gabbouj
Abstract: Deep learning based compressive sensing (CS) methods typically learn sampling operators using convolutional or block wise fully connected layers, which limit receptive fields and scale poorly for high dimensional data. We propose MTSCSNet, a CS framework based on Multiscale Tensor Summation (MTS) factorization, a structured operator for efficient multidimensional signal processing. MTS performs mode-wise linear transformations with multiscale summation, enabling large receptive fields and effective modeling of cross-dimensional correlations. In MTSCSNet, MTS is first used as a learnable CS operator that performs linear dimensionality reduction in tensor space, with its adjoint defining the initial back-projection, and is then applied in the reconstruction stage to directly refine this estimate. This results in a simple feed-forward architecture without iterative or proximal optimization, while remaining parameter and computation efficient. Experiments on standard CS benchmarks show that MTSCSNet achieves state-of-the-art reconstruction performance on RGB images, with notable PSNR gains and faster inference, even compared to recent diffusion-based CS methods, while using a significantly more compact feed-forward architecture.
Authors: Carolina R. Kelsch, Leonardo S. B. Pereira, Natnael Mola, Luis H. Arribas, Juan C. S. M. Avedillo
Abstract: Machine Unlearning aims to remove the influence of specific data or concepts from trained models while preserving overall performance, a capability increasingly required by data protection regulations and responsible AI practices. Despite recent progress, unlearning in text-to-image diffusion models remains challenging due to high computational costs and the difficulty of balancing effective forgetting with retention of unrelated concepts. We introduce FADE (Fast Adapter for Data Erasure), a two-stage unlearning method for image generation that combines parameter localization with self-distillation. FADE first identifies parameters most responsible for the forget set using gradient-based saliency and constrains updates through sparse LoRA adapters, ensuring lightweight, localized modifications. In a second stage, FADE applies a self-distillation objective that overwrites the forgotten concept with a user-defined surrogate while preserving behavior on retained data. The resulting adapters are memory-efficient, reversible, and can be merged or removed at runtime, enabling flexible deployment in production systems. We evaluated FADE on the UnlearnCanvas benchmark and conducted ablation studies on Imagenette, Labeled Faces in the Wild, AtharvaTaras Dog Breeds Dataset, and SUN Attributes datasets, demonstrating State-of-the-Art unlearning performance with fine-grained control over the forgetting-retention trade-off. Our results demonstrate that FADE achieves strong concept erasure and high retainability across various domains, making it a suitable solution for selective unlearning in diffusion-based image generation models.
Authors: Francesca Da Ros, Tarik Za\v{c}iragi\'c, Aske Plaat, Thomas B\"ack, Niki van Stein
Abstract: Reproducibility is an important requirement in evolutionary computation, where results largely depend on computational experiments. In practice, reproducibility relies on how algorithms, experimental protocols, and artifacts are documented and shared. Despite growing awareness, there is still limited empirical evidence on the actual reproducibility levels of published work in the field. In this paper, we study the reproducibility practices in papers published in the Evolutionary Combinatorial Optimization and Metaheuristics track of the Genetic and Evolutionary Computation Conference over a ten-year period. We introduce a structured reproducibility checklist and apply it through a systematic manual assessment of the selected corpus. In addition, we propose RECAP (REproducibility Checklist Automation Pipeline), an LLM-based system that automatically evaluates reproducibility signals from paper text and associated code repositories. Our analysis shows that papers achieve an average completeness score of 0.62, and that 36.90% of them provide additional material beyond the manuscript itself. We demonstrate that automated assessment is feasible: RECAP achieves substantial agreement with human evaluators (Cohen's k of 0.67). Together, these results highlight persistent gaps in reproducibility reporting and suggest that automated tools can effectively support large-scale, systematic monitoring of reproducibility practices.
Authors: Daniel Nobrega
Abstract: We present TACIT (Transformation-Aware Capturing of Implicit Thought), a diffusion-based transformer for interpretable visual reasoning. Unlike language-based reasoning systems, TACIT operates entirely in pixel space using rectified flow, enabling direct visualization of the reasoning process at each inference step. We demonstrate the approach on maze-solving, where the model learns to transform images of unsolved mazes into solutions. Key results on 1 million synthetic maze pairs include: - 192x reduction in training loss over 100 epochs - 22.7x improvement in L2 distance to ground truth - Only 10 Euler steps required (vs. 100-1000 for typical diffusion models) Quantitative analysis reveals a striking phase transition phenomenon: the solution remains invisible for 68% of the transformation (zero recall), then emerges abruptly at t=0.70 within just 2% of the process. Most remarkably, 100% of samples exhibit simultaneous emergence across all spatial regions, ruling out sequential path construction and providing evidence for holistic rather than algorithmic reasoning. This "eureka moment" pattern -- long incubation followed by sudden crystallization -- parallels insight phenomena in human cognition. The pixel-space design with noise-free flow matching provides a foundation for understanding how neural networks develop implicit reasoning strategies that operate below and before language.
Authors: Serkan Sulun
Abstract: As the volume of video content on the internet grows rapidly, finding a suitable soundtrack remains a significant challenge. This thesis presents EMSYNC (EMotion and SYNChronization), a fast, free, and automatic solution that generates music tailored to the input video, enabling content creators to enhance their productions without composing or licensing music. Our model creates music that is emotionally and rhythmically synchronized with the video. A core component of EMSYNC is a novel video emotion classifier. By leveraging pretrained deep neural networks for feature extraction and keeping them frozen while training only fusion layers, we reduce computational complexity while improving accuracy. We show the generalization abilities of our method by obtaining state-of-the-art results on Ekman-6 and MovieNet. Another key contribution is a large-scale, emotion-labeled MIDI dataset for affective music generation. We then present an emotion-based MIDI generator, the first to condition on continuous emotional values rather than discrete categories, enabling nuanced music generation aligned with complex emotional content. To enhance temporal synchronization, we introduce a novel temporal boundary conditioning method, called "boundary offset encodings," aligning musical chords with scene changes. Combining video emotion classification, emotion-based music generation, and temporal boundary conditioning, EMSYNC emerges as a fully automatic video-based music generator. User studies show that it consistently outperforms existing methods in terms of music richness, emotional alignment, temporal synchronization, and overall preference, setting a new state-of-the-art in video-based music generation.
Authors: Ali Alqutayfi, Sadam Al-Azani
Abstract: MRI cross-modal synthesis involves generating images from one acquisition protocol using another, offering considerable clinical value by reducing scan time while maintaining diagnostic information. This paper presents a comprehensive comparison of three state-of-the-art generative models for T1-to-T2 MRI reconstruction: Pix2Pix GAN, CycleGAN, and Variational Autoencoder (VAE). Using the BraTS 2020 dataset (11,439 training and 2,000 testing slices), we evaluate these models based on established metrics including Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index (SSIM). Our experiments demonstrate that all models can successfully synthesize T2 images from T1 inputs, with CycleGAN achieving the highest PSNR (32.28 dB) and SSIM (0.9008), while Pix2Pix GAN provides the lowest MSE (0.005846). The VAE, though showing lower quantitative performance (MSE: 0.006949, PSNR: 24.95 dB, SSIM: 0.6573), offers advantages in latent space representation and sampling capabilities. This comparative study provides valuable insights for researchers and clinicians selecting appropriate generative models for MRI synthesis applications based on their specific requirements and data constraints.
Authors: Zihao Fan, Xin Lu, Yidi Liu, Jie Huang, Dong Li, Xueyang Fu, Zheng-Jun Zha
Abstract: Diffusion-based super-resolution can synthesize rich details, but models trained on synthetic paired data often fail on real-world LR images due to distribution shifts. We propose Bird-SR, a bidirectional reward-guided diffusion framework that formulates super-resolution as trajectory-level preference optimization via reward feedback learning (ReFL), jointly leveraging synthetic LR-HR pairs and real-world LR images. For structural fidelity easily affected in ReFL, the model is directly optimized on synthetic pairs at early diffusion steps, which also facilitates structure preservation for real-world inputs under smaller distribution gap in structure levels. For perceptual enhancement, quality-guided rewards are applied at later sampling steps to both synthetic and real LR images. To mitigate reward hacking, the rewards for synthetic results are formulated in a relative advantage space bounded by their clean counterparts, while real-world optimization is regularized via a semantic alignment constraint. Furthermore, to balance structural and perceptual learning, we adopt a dynamic fidelity-perception weighting strategy that emphasizes structure preservation at early stages and progressively shifts focus toward perceptual optimization at later diffusion steps. Extensive experiments on real-world SR benchmarks demonstrate that Bird-SR consistently outperforms state-of-the-art methods in perceptual quality while preserving structural consistency, validating its effectiveness for real-world super-resolution.
Authors: S M Rakib UI Karim, Wenyi Lu, Sean Goggins
Abstract: Open-source software (OSS) is foundational to modern digital infrastructure, yet this context for group work continues to struggle to ensure sufficient contributions in many critical cases. This literature review explores how artificial intelligence (AI) is being leveraged to address critical challenges to OSS sustainability, including maintaining contributor engagement, securing funding, ensuring code quality and security, fostering healthy community dynamics, and preventing project abandonment. Synthesizing recent interdisciplinary research, the paper identifies key applications of AI in this domain, including automated bug triaging, system maintenance, contributor onboarding and mentorship, community health analytics, vulnerability detection, and task automation. The review also examines the limitations and ethical concerns that arise from applying AI in OSS contexts, including data availability, bias and fairness, transparency, risks of misuse, and the preservation of human-centered values in collaborative development. By framing AI not as a replacement but as a tool to augment human infrastructure, this study highlights both the promise and pitfalls of AI-driven interventions. It concludes by identifying critical research gaps and proposing future directions at the intersection of AI, sustainability, and OSS, aiming to support more resilient and equitable open-source ecosystems.
Authors: Nardine Basta, Firas Ben Hmida, Houssem Jmal, Muhammad Ikram, Mohamed Ali Kaafar, Andy Walker
Abstract: In today's enterprise network landscape, the combination of perimeter and distributed firewall rules governs connectivity. To address challenges arising from increased traffic and diverse network architectures, organizations employ automated tools for firewall rule and access policy generation. Yet, effectively managing risks arising from dynamically generated policies, especially concerning critical asset exposure, remains a major challenge. This challenge is amplified by evolving network structures due to trends like remote users, bring-your-own devices, and cloud integration. This paper introduces a novel graph neural network model for identifying weighted shortest paths. The model aids in detecting network misconfigurations and high-risk connectivity paths that threaten critical assets, potentially exploited in zero-day attacks -- cyber-attacks exploiting undisclosed vulnerabilities. The proposed Pro-ZD framework adopts a proactive approach, automatically fine-tuning firewall rules and access policies to address high-risk connections and prevent unauthorized access. Experimental results highlight the robustness and transferability of Pro-ZD, achieving over 95% average accuracy in detecting high-risk connections. \
Authors: Xinwu Ye, Yicheng Mao, Jia Zhang, Yimeng Liu, Li Hao, Fang Wu, Zhiwei Li, Yuxuan Liao, Zehong Wang, Zhiyuan Liu, Zhenfei Yin, Li Yuan, Philip Torr, Huan Sun, Xiangxiang Zeng, Mengdi Wang, Le Cong, Shenghua Gao, Xiangru Tang
Abstract: Chemical large language models (LLMs) predominantly rely on explicit Chain-of-Thought (CoT) in natural language to perform complex reasoning. However, chemical reasoning is inherently continuous and structural, and forcing it into discrete linguistic tokens introduces a fundamental representation mismatch that constrains both efficiency and performance. We introduce LatentChem, a latent reasoning interface that decouples chemical computation from textual generation, enabling models to perform multi-step reasoning directly in continuous latent space while emitting language only for final outputs. Remarkably, we observe a consistent emergent behavior: when optimized solely for task success, models spontaneously internalize reasoning, progressively abandoning verbose textual derivations in favor of implicit latent computation. This shift is not merely stylistic but computationally advantageous. Across diverse chemical reasoning benchmarks, LatentChem achieves a 59.88\% non-tie win rate over strong CoT-based baselines on ChemCoTBench, while delivering a 10.84$\times$ average inference speedup. Our results provide empirical evidence that chemical reasoning is more naturally and effectively realized as continuous latent dynamics rather than discretized linguistic trajectories.
Authors: Videet Mehta, Liming Wang, Hilde Kuehne, Rogerio Feris, James R. Glass, M. Jehanzeb Mirza
Abstract: Large audio-language models (LALMs) exhibit strong zero-shot capabilities in multiple downstream tasks, such as audio question answering (AQA) and abstract reasoning; however, these models still lag behind specialized models for certain discriminative tasks (e.g., audio classification). Recent studies show that sparse subsets of attention heads within an LALM can serve as strong discriminative feature extractors for downstream tasks such as classification via simple voting schemes. However, these methods assign uniform weights to all selected heads, implicitly assuming that each head contributes equally across all semantic categories. In this work, we propose Class-Conditional Sparse Attention Vectors for Large Audio-Language Models, a few-shot classification method that learns class-dependent importance weights over attention heads. This formulation allows individual heads to specialize in distinct semantic categories and to contribute to ensemble predictions proportionally to their estimated reliability. Experiments on multiple few-shot audio and audiovisual classification benchmarks and tasks demonstrate that our method consistently outperforms state-of-the-art uniform voting-based approaches by up to 14.52%, 1.53%, 8.35% absolute gains for audio classification, audio-visual classification, and spoofing detection respectively.
Authors: Yingru Li, Jiawei Xu, Ziniu Li, Jiacai Liu, Wei Liu, Yuxuan Tong, Longtao Zheng, Zhenghai Xue, Yaxiang Zhang, Tianle Cai, Ge Zhang, Qian Liu, Baoxiang Wang
Abstract: Reinforcement Learning (RL) for Large Language Models (LLMs) often suffers from training collapse in long-horizon tasks due to exploding gradient variance. To mitigate this, a baseline is commonly introduced for advantage computation; however, traditional value models remain difficult to optimize, and standard group-based baselines overlook sequence heterogeneity. Although classic optimal baseline theory can achieve global variance reduction, it neglects token heterogeneity and requires prohibitive gradient-based computation. In this work, we derive the Optimal Token Baseline (OTB) from first principles, proving that gradient updates should be weighted inversely to their cumulative gradient norm. To ensure efficiency, we propose the Logit-Gradient Proxy that approximates the gradient norm using only forward-pass probabilities. Our method achieves training stability and matches the performance of large group sizes ($N=32$) with only $N=4$, reducing token consumption by over 65% across single-turn and tool-integrated reasoning tasks.
Authors: Yicheng He, Zheng Zhao, Zhou Kaiyu, Bryan Dai, Jie Fu, Yonghui Yang
Abstract: Current paradigms for code verification rely heavily on external mechanisms-such as execution-based unit tests or auxiliary LLM judges-which are often labor-intensive or limited by the judging model's own capabilities. This raises a fundamental, yet unexplored question: Can an LLM's functional correctness be assessed purely from its internal computational structure? Our primary objective is to investigate whether the model's neural dynamics encode internally decodable signals that are predictive of logical validity during code generation. Inspired by mechanistic interpretability, we propose to treat code verification as a mechanistic diagnostic task, mapping the model's explicit algorithmic trajectory into line-level attribution graphs. By decomposing complex residual flows, we aim to identify the structural signatures that distinguish sound reasoning from logical failure within the model's internal circuits. Analysis across Python, C++, and Java confirms that intrinsic correctness signals are robust across diverse syntaxes. Topological features from these internal graphs predict correctness more reliably than surface heuristics and enable targeted causal interventions to fix erroneous logic. These findings establish internal introspection as a decodable property for verifying generated code. Our code is at https:// github.com/bruno686/CodeCircuit.
Authors: Thu Hang Phung, Duong M. Nguyen, Thanh Trung Huynh, Quoc Viet Hung Nguyen, Trong Nghia Hoang, Phi Le Nguyen
Abstract: This paper introduces a generalized federated prompt-tuning framework for practical scenarios where local datasets are multi-modal and exhibit different distributional patterns of missing features at the input level. The proposed framework bridges the gap between federated learning and multi-modal prompt-tuning which have traditionally focused on either uni-modal or centralized data. A key challenge in this setting arises from the lack of semantic alignment between prompt instructions that encode similar distributional patterns of missing data across different clients. To address this, our framework introduces specialized client-tuning and server-aggregation designs that simultaneously optimize, align, and aggregate prompt-tuning instructions across clients and data modalities. This allows prompt instructions to complement one another and be combined effectively. Extensive evaluations on diverse multimodal benchmark datasets demonstrate that our work consistently outperforms state-of-the-art (SOTA) baselines.
Authors: Haoming Wang, Qiyao Xue, Weichen Liu, Wei Gao
Abstract: When embodied AI is expanding from traditional object detection and recognition to more advanced tasks of robot manipulation and actuation planning, visual spatial reasoning from the video inputs is necessary to perceive the spatial relationships of objects and guide device actions. However, existing visual language models (VLMs) have very weak capabilities in spatial reasoning due to the lack of knowledge about 3D spatial information, especially when the reasoning task involve complex spatial relations across multiple video frames. In this paper, we present a new inference-time computing technique for on-device embodied AI, namely \emph{MosaicThinker}, which enhances the on-device small VLM's spatial reasoning capabilities on difficult cross-frame reasoning tasks. Our basic idea is to integrate fragmented spatial information from multiple frames into a unified space representation of global semantic map, and further guide the VLM's spatial reasoning over the semantic map via a visual prompt. Experiment results show that our technique can greatly enhance the accuracy of cross-frame spatial reasoning on resource-constrained embodied AI devices, over reasoning tasks with diverse types and complexities.
Authors: Yongqing Jiang, Jianze Wang, Zhiqi Shen, Zhenghong Lin, Jiayuan Wang, Yijian Yang, Kaoshan Dai, Haoran Luo
Abstract: Structural modeling is a fundamental component of computational engineering science, in which even minor physical inconsistencies or specification violations may invalidate downstream simulations. The potential of large language models (LLMs) for automatic generation of modeling code has been demonstrated. However, non-executable or physically inconsistent outputs remain prevalent under stringent engineering constraints. A framework for physics-consistent automatic building modeling is therefore proposed, integrating domain knowledge construction, constraint-oriented model alignment, and verification-driven evaluation. CivilInstruct is introduced as a domain-specific dataset that formalizes structural engineering knowledge and constraint reasoning to enable simulation-ready model generation. A two-stage fine-tuning strategy is further employed to enforce constraint satisfaction and application programming interface compliance, substantially reducing hallucinated and non-conforming outputs. MBEval is presented as a verification-driven benchmark that evaluates executability and structural dynamics consistency through closed-loop validation. Experimental results show consistent improvements over baselines across rigorous verification metrics. Our code is available at https://github.com/Jovanqing/AutoBM.
Authors: Wenda Wang, Yang Zhang, Zhewei Wei, Wenbing Huang
Abstract: Antigen-antibody binding is a critical process in the immune response. Although recent progress has advanced antibody design, current methods lack a generative framework for end-to-end modeling of full-atom antibody structures and struggle to fully exploit antigen-specific geometric information for optimizing local binding interfaces and global structures. To overcome these limitations, we introduce AbFlow, a flow-matching framework that leverages optimal transport to design full-atom antibodies end-to-end. AbFlow incorporates an extended velocity field network featuring an equivariant Surface Multi-channel Encoder, which uses surface-level antigen interaction data to refine the antibody structure, particularly the CDR-H3 region. Extensive experiments in paratoep-centric antibody design, multi-CDRs and full-atom antibody design, binding affinity optimization, and complex structure prediction show that AbFlow produces superior antigen-antibody complexes, especially at the contact interface, and markedly improves the binding affinity of generated antibodies.
Authors: Jun Han, Shuo Zhang, Wei Li, Zhi Yang, Yifan Dong, Tu Hu, Jialuo Yuan, Xiaomin Yu, Yumo Zhu, Fangqi Lou, Xin Guo, Zhaowei Liu, Tianyi Jiang, Ruichuan An, Jingping Liu, Biao Wu, Rongze Chen, Kunyi Wang, Yifan Wang, Sen Hu, Xinbing Kong, Liwen Zhang, Ronghao Chen, Huacan Wang
Abstract: Financial markets are noisy and non-stationary, making alpha mining highly sensitive to noise in backtesting results and sudden market regime shifts. While recent agentic frameworks improve alpha mining automation, they often lack controllable multi-round search and reliable reuse of validated experience. To address these challenges, we propose QuantaAlpha, an evolutionary alpha mining framework that treats each end-to-end mining run as a trajectory and improves factors through trajectory-level mutation and crossover operations. QuantaAlpha localizes suboptimal steps in each trajectory for targeted revision and recombines complementary high-reward segments to reuse effective patterns, enabling structured exploration and refinement across mining iterations. During factor generation, QuantaAlpha enforces semantic consistency across the hypothesis, factor expression, and executable code, while constraining the complexity and redundancy of the generated factor to mitigate crowding. Extensive experiments on the China Securities Index 300 (CSI 300) demonstrate consistent gains over strong baseline models and prior agentic systems. When utilizing GPT-5.2, QuantaAlpha achieves an Information Coefficient (IC) of 0.1501, with an Annualized Rate of Return (ARR) of 27.75% and a Maximum Drawdown (MDD) of 7.98%. Moreover, factors mined on CSI 300 transfer effectively to the China Securities Index 500 (CSI 500) and the Standard & Poor's 500 Index (S&P 500), delivering 160% and 137% cumulative excess return over four years, respectively, which indicates strong robustness of QuantaAlpha under market distribution shifts.
Authors: Michael Marketsm\"uller, Simon Martin, Tim Schlippe
Abstract: Enterprise systems increasingly require natural language interfaces that can translate user requests into structured operations such as SQL queries and REST API calls. While large language models (LLMs) show promise for code generation [Chen et al., 2021; Huynh and Lin, 2025], their effectiveness in domain-specific enterprise contexts remains underexplored, particularly when both retrieval and modification tasks must be handled jointly. This paper presents a comprehensive evaluation of three retrieval-augmented generation (RAG) variants [Lewis et al., 2021] -- standard RAG, Self-RAG [Asai et al., 2024], and CoRAG [Wang et al., 2025] -- across SQL query generation, REST API call generation, and a combined task requiring dynamic task classification. Using SAP Transactional Banking as a realistic enterprise use case, we construct a novel test dataset covering both modalities and evaluate 18 experimental configurations under database-only, API-only, and hybrid documentation contexts. Results demonstrate that RAG is essential: Without retrieval, exact match accuracy is 0% across all tasks, whereas retrieval yields substantial gains in execution accuracy (up to 79.30%) and component match accuracy (up to 78.86%). Critically, CoRAG proves most robust in hybrid documentation settings, achieving statistically significant improvements in the combined task (10.29% exact match vs. 7.45% for standard RAG), driven primarily by superior SQL generation performance (15.32% vs. 11.56%). Our findings establish retrieval-policy design as a key determinant of production-grade natural language interfaces, showing that iterative query decomposition outperforms both top-k retrieval and binary relevance filtering under documentation heterogeneity.
Authors: Gyoung S. Na, Chanyoung Park
Abstract: Various representation learning methods for molecular structures have been devised to accelerate data-driven chemistry. However, the representation capabilities of existing methods are essentially limited to atom-level information, which is not sufficient to describe real-world molecular physics. Although electron-level information can provide fundamental knowledge about chemical compounds beyond the atom-level information, obtaining the electron-level information in real-world molecules is computationally impractical and sometimes infeasible. We propose a method for learning electron-informed molecular representations without additional computation costs by transferring readily accessible electron-level information about small molecules to large molecules of our interest. The proposed method achieved state-of-the-art prediction accuracy on extensive benchmark datasets containing experimentally observed molecular physics. The source code for HEDMoL is available at https://github.com/ngs00/HEDMoL.
Authors: Yu-Che Tsai, Hsiang Hsiao, Kuan-Yu Chen, Shou-De Lin
Abstract: Text embeddings enable numerous NLP applications but face severe privacy risks from embedding inversion attacks, which can expose sensitive attributes or reconstruct raw text. Existing differential privacy defenses assume uniform sensitivity across embedding dimensions, leading to excessive noise and degraded utility. We propose SPARSE, a user-centric framework for concept-specific privacy protection in text embeddings. SPARSE combines (1) differentiable mask learning to identify privacy-sensitive dimensions for user-defined concepts, and (2) the Mahalanobis mechanism that applies elliptical noise calibrated by dimension sensitivity. Unlike traditional spherical noise injection, SPARSE selectively perturbs privacy-sensitive dimensions while preserving non-sensitive semantics. Evaluated across six datasets with three embedding models and attack scenarios, SPARSE consistently reduces privacy leakage while achieving superior downstream performance compared to state-of-the-art DP methods.
Authors: Haipeng Jiang, Kailong Ren, Zimo Yin, Zhetao Sun, Xin Gan, Guangyi Lv, Ming He, Peng Wang, Congli Yin, Hong Pan, Changwen Zhang, Shan Tong, Zhengyu Xu, Zeping Chen, Yubin Huangfu, Yanzhi Xu, Xing Su, Qin Feng, Dong An, Jianping Fan
Abstract: Recent advanced LLM-powered agent systems have exhibited their remarkable capabilities in tackling complex, long-horizon tasks. Nevertheless, they still suffer from inherent limitations in resource efficiency, context management, and multimodal perception. Based on these observations, Lemon Agent is introduced, a multi-agent orchestrator-worker system built on a newly proposed AgentCortex framework, which formalizes the classic Planner-Executor-Memory paradigm through an adaptive task execution mechanism. Our system integrates a hierarchical self-adaptive scheduling mechanism that operates at both the overall orchestrator layer and workers layer. This mechanism can dynamically adjust computational intensity based on task complexity. It enables orchestrator to allocate one or more workers for parallel subtask execution, while workers can further improve operational efficiency by invoking tools concurrently. By virtue of this two-tier architecture, the system achieves synergistic balance between global task coordination and local task execution, thereby optimizing resource utilization and task processing efficiency in complex scenarios. To reduce context redundancy and increase information density during parallel steps, we adopt a three-tier progressive context management strategy. To make fuller use of historical information, we propose a self-evolving memory system, which can extract multi-dimensional valid information from all historical experiences to assist in completing similar tasks. Furthermore, we provide an enhanced MCP toolset. Empirical evaluations on authoritative benchmarks demonstrate that our Lemon Agent can achieve a state-of-the-art 91.36% overall accuracy on GAIA and secures the top position on the xbench-DeepSearch leaderboard with a score of 77+.
Authors: Wang Lin, Feng Wang, Majun Zhang, Wentao Hu, Tao Jin, Zhou Zhao, Fei Wu, Jingyuan Chen, Alan Yuille, Sucheng Ren
Abstract: Recent advances in image editing models have demonstrated remarkable capabilities in executing explicit instructions, such as attribute manipulation, style transfer, and pose synthesis. However, these models often face challenges when dealing with implicit editing instructions, which describe the cause of a visual change without explicitly detailing the resulting outcome. These limitations arise because existing models rely on uniform editing strategies that are not equipped to handle the complex world knowledge and reasoning required for implicit instructions. To address this gap, we introduce \textbf{WorldEdit}, a dataset specifically designed to enable world-driven image editing. WorldEdit consists of high-quality editing samples, guided by paraphrased instructions that align with real-world causal logic. Furthermore, we provide \textbf{WorldEdit-Test} for evaluating the existing model's performance on causal editing scenarios. With WorldEdit, we use a two-stage training framework for fine-tuning models like Bagel, integrating with a causal verification reward. Our results show that the proposed dataset and methods significantly narrow the gap with GPT-4o and Nano-Banana, demonstrating competitive performance not only in instruction following but also in knowledge plausibility, where many open-source systems typically struggle.
Authors: Yuyang Dai, Yan Lin, Zhuohan Xie, Yuxia Wang
Abstract: Reliable financial reasoning requires knowing not only how to answer, but also when an answer cannot be justified. In real financial practice, problems often rely on implicit assumptions that are taken for granted rather than stated explicitly, causing problems to appear solvable while lacking enough information for a definite answer. We introduce REALFIN, a bilingual benchmark that evaluates financial reasoning by systematically removing essential premises from exam-style questions while keeping them linguistically plausible. Based on this, we evaluate models under three formulations that test answering, recognizing missing information, and rejecting unjustified options, and find consistent performance drops when key conditions are absent. General-purpose models tend to over-commit and guess, while most finance-specialized models fail to clearly identify missing premises. These results highlight a critical gap in current evaluations and show that reliable financial models must know when a question should not be answered.
Authors: L\'eon Zheng (MBZUAI, LRE), Thomas Hirtz (MBZUAI, LRE), Yazid Janati (MBZUAI, LRE), Eric Moulines (MBZUAI, LRE)
Abstract: Zero-shot diffusion posterior sampling offers a flexible framework for inverse problems by accommodating arbitrary degradation operators at test time, but incurs high computational cost due to repeated likelihood-guided updates. In contrast, previous amortized diffusion approaches enable fast inference by replacing likelihood-based sampling with implicit inference models, but at the expense of robustness to unseen degradations. We introduce an amortization strategy for diffusion posterior sampling that preserves explicit likelihood guidance by amortizing the inner optimization problems arising in variational diffusion posterior sampling. This accelerates inference for in-distribution degradations while maintaining robustness to previously unseen operators, thereby improving the trade-off between efficiency and flexibility in diffusion-based inverse problems.
Authors: Chenglei Yu, Chuanrui Wang, Bangyan Liao, Tailin Wu
Abstract: A central goal in systems biology and drug discovery is to predict the transcriptional response of cells to perturbations. This task is challenging due to the noisy and sparse nature of single-cell measurements, as well as the fact that perturbations often induce population-level shifts rather than changes in individual cells. Existing deep learning methods typically assume cell-level correspondences, limiting their ability to capture such global effects. We present scDFM, a generative framework based on conditional flow matching that models the full distribution of perturbed cells conditioned on control states. By incorporating a maximum mean discrepancy (MMD) objective, our method aligns perturbed and control populations beyond cell-level correspondences. To further improve robustness to sparsity and noise, we introduce the Perturbation-Aware Differential Transformer (PAD-Transformer), a backbone architecture that leverages gene interaction graphs and differential attention to capture context-specific expression changes. Across multiple genetic and drug perturbation benchmarks, scDFM consistently outperforms prior methods, demonstrating strong generalization in both unseen and combinatorial settings. In the combinatorial setting, it reduces mean squared error by 19.6% relative to the strongest baseline. These results highlight the importance of distribution-level generative modeling for robust in silico perturbation prediction. The code is available at https://github.com/AI4Science-WestlakeU/scDFM
Authors: Zhuoheng Li, Ying Chen
Abstract: Multimodal large language models (MLLMs) have advanced the capabilities to interpret and act on visual input in 3D environments, empowering diverse applications such as robotics and situated conversational agents. When MLLMs reason over camera-captured views of the physical world, a new attack surface emerges: an attacker can place text-bearing physical objects in the environment to override MLLMs' intended task. While prior work has studied prompt injection in the text domain and through digitally edited 2D images, it remains unclear how these attacks function in 3D physical environments. To bridge the gap, we introduce PI3D, a prompt injection attack against MLLMs in 3D environments, realized through text-bearing physical object placement rather than digital image edits. We formulate and solve the problem of identifying an effective 3D object pose (position and orientation) with injected text, where the attacker's goal is to induce the MLLM to perform the injected task while ensuring that the object placement remains physically plausible. Experiments demonstrate that PI3D is an effective attack against multiple MLLMs under diverse camera trajectories. We further evaluate existing defenses and show that they are insufficient to defend against PI3D.
Authors: Haoyu Zhang, Zhipeng Li, Yiwen Guo, Tianshu Yu
Abstract: Omni-modal large language models (OLLMs) aim to unify multimodal understanding and generation, yet incorporating speech with 3D facial animation remains largely unexplored despite its importance for natural interaction. A key challenge arises from the representation mismatch between discrete, token-level semantic reasoning in LLMs and the dense, fine-grained temporal dynamics required for 3D facial motion, which makes direct modeling difficult to optimize under limited data. We propose Expressive Omni (Ex-Omni), an open-source omni-modal framework that augments OLLMs with speech-accompanied 3D facial animation. Ex-Omni reduces learning difficulty by decoupling semantic reasoning from temporal generation, leveraging speech units as temporal scaffolding and a unified token-as-query gated fusion (TQGF) mechanism for controlled semantic injection. We further introduce InstructEx, a dataset aims to facilitate augment OLLMs with speech-accompanied 3D facial animation. Extensive experiments demonstrate that Ex-Omni performs competitively against existing open-source OLLMs while enabling stable aligned speech and facial animation generation.
Authors: Shang Liu, Hanyu Pei, Zeyan Liu
Abstract: Large Language Models(LLMs) have been successful in numerous fields. Alignment has usually been applied to prevent them from harmful purposes. However, aligned LLMs remain vulnerable to jailbreak attacks that deliberately mislead them into producing harmful outputs. Existing jailbreaks are either black-box, using carefully crafted, unstealthy prompts, or white-box, requiring resource-intensive computation. In light of these challenges, we introduce ShallowJail, a novel attack that exploits shallow alignment in LLMs. ShallowJail can misguide LLMs' responses by manipulating the initial tokens during inference. Through extensive experiments, we demonstrate the effectiveness of~\shallow, which substantially degrades the safety of state-of-the-art LLM responses.
Authors: Jianrui Zhang, Anirudh Sundara Rajan, Brandon Han, Soochahn Lee, Sukanta Ganguly, Yong Jae Lee
Abstract: Universal Multimodal Retrieval (UMR) seeks any-to-any search across text and vision, yet modern embedding models remain brittle when queries require latent reasoning (e.g., resolving underspecified references or matching compositional constraints). We argue this brittleness is often data-induced: when images carry "silent" evidence and queries leave key semantics implicit, a single embedding pass must both reason and compress, encouraging spurious feature matching. We propose a data-centric framework that decouples these roles by externalizing reasoning before retrieval. Using a strong Vision--Language Model, we make implicit semantics explicit by densely captioning visual evidence in corpus entries, resolving ambiguous multimodal references in queries, and rewriting verbose instructions into concise retrieval constraints. Inference-time enhancement alone is insufficient; the retriever must be trained on these semantically dense representations to avoid distribution shift and fully exploit the added signal. Across M-BEIR, our reasoning-augmented training method yields consistent gains over strong baselines, with ablations showing that corpus enhancement chiefly benefits knowledge-intensive queries while query enhancement is critical for compositional modification requests. We publicly release our code at https://github.com/AugmentedRetrieval/ReasoningAugmentedRetrieval.
URLs: https://github.com/AugmentedRetrieval/ReasoningAugmentedRetrieval.
Authors: Jiaqing Chen, Nicholas Hadler, Tiankai Xie, Rostyslav Hnatyshyn, Caleb Geniesse, Yaoqing Yang, Michael W. Mahoney, Talita Perciano, John F. Hartwig, Ross Maciejewski, Gunther H. Weber
Abstract: Loss landscapes are a powerful tool for understanding neural network optimization and generalization, yet traditional low-dimensional analyses often miss complex topological features. We present Landscaper, an open-source Python package for arbitrary-dimensional loss landscape analysis. Landscaper combines Hessian-based subspace construction with topological data analysis to reveal geometric structures such as basin hierarchy and connectivity. A key component is the Saddle-Minimum Average Distance (SMAD) for quantifying landscape smoothness. We demonstrate Landscaper's effectiveness across various architectures and tasks, including those involving pre-trained language models, showing that SMAD captures training transitions, such as landscape simplification, that conventional metrics miss. We also illustrate Landscaper's performance in challenging chemical property prediction tasks, where SMAD can serve as a metric for out-of-distribution generalization, offering valuable insights for model diagnostics and architecture design in data-scarce scientific machine learning scenarios.
Authors: Prerna Ravi, Car\'umey Stevens, Beatriz Flamia Azevedo, Jasmine David, Brandon Hanks, Hal Abelson, Grace Lin, Emma Anderson
Abstract: Collaboration is a cornerstone of 21st-century learning, yet teachers continue to face challenges in supporting productive peer interaction. Emerging generative AI tools offer new possibilities for scaffolding collaboration, but their role in mediating in-person group work remains underexplored, especially from the perspective of educators. This paper presents findings from an exploratory qualitative study with 33 K12 teachers who interacted with Phoenix, a voice-based conversational agent designed to function as a near-peer in face-to-face group collaboration. Drawing on playtesting sessions, surveys, and focus groups, we examine how teachers perceived the agent's behavior, its influence on group dynamics, and its classroom potential. While many appreciated Phoenix's capacity to stimulate engagement, they also expressed concerns around autonomy, trust, anthropomorphism, and pedagogical alignment. We contribute empirical insights into teachers' mental models of AI, reveal core design tensions, and outline considerations for group-facing AI agents that support meaningful, collaborative learning.
Authors: Samuel Daulton, David Eriksson, Maximilian Balandat, Eytan Bakshy
Abstract: Bayesian optimization (BO) is a popular technique for sample-efficient optimization of black-box functions. In many applications, the parameters being tuned come with a carefully engineered default configuration, and practitioners only want to deviate from this default when necessary. Standard BO, however, does not aim to minimize deviation from the default and, in practice, often pushes weakly relevant parameters to the boundary of the search space. This makes it difficult to distinguish between important and spurious changes and increases the burden of vetting recommendations when the optimization objective omits relevant operational considerations. We introduce BONSAI, a default-aware BO policy that prunes low-impact deviations from a default configuration while explicitly controlling the loss in acquisition value. BONSAI is compatible with a variety of acquisition functions, including expected improvement and upper confidence bound (GP-UCB). We theoretically bound the regret incurred by BONSAI, showing that, under certain conditions, it enjoys the same no-regret property as vanilla GP-UCB. Across many real-world applications, we empirically find that BONSAI substantially reduces the number of non-default parameters in recommended configurations while maintaining competitive optimization performance, with little effect on wall time.
Authors: Bjarni Haukur Bjarnason, Andr\'e Silva, Martin Monperrus
Abstract: Agentic systems are evaluated on benchmarks where agents interact with environments to solve tasks. Most papers report a pass@1 score computed from a single run per task, assuming this gives a reliable performance estimate. We test this assumption by collecting 60,000 agentic trajectories on SWE-Bench-Verified, spanning three models and two scaffolds. We find substantial variance: single-run pass@1 estimates vary by 2.2 to 6.0 percentage points depending on which run is selected, with standard deviations exceeding 1.5 percentage points even at temperature 0. This variance has critical implications: reported improvements of 2--3 percentage points may reflect evaluation noise rather than genuine algorithmic progress. Through token-level analysis, we show that trajectories diverge early, often within the first few percent of tokens, and that these small differences cascade into different solution strategies. To enable reliable evaluation of agentic systems, we recommend three concrete practices: (1) estimate pass@1 from multiple independent runs per task, especially when measuring small improvements, (2) use statistical power analysis to determine the number of runs needed to detect expected effect sizes, and (3) consider metrics like pass@k (optimistic bound) and pass^k (pessimistic bound) with k>1 to better characterize the full performance envelope. While these practices increase evaluation cost, they are essential for distinguishing genuine scientific progress from statistical noise.
Authors: Ayush Roy, Rudrasis Chakraborty, Lav Varshney, Vishnu Suresh Lokhande
Abstract: Pooling heterogeneous datasets across domains is a common strategy in representation learning, but naive pooling can amplify distributional asymmetries and yield biased estimators, especially in settings where zero-shot generalization is required. We propose a matching framework that selects samples relative to an adaptive centroid and iteratively refines the representation distribution. The double robustness and the propensity score matching for the inclusion of data domains make matching more robust than naive pooling and uniform subsampling by filtering out the confounding domains (the main cause of heterogeneity). Theoretical and empirical analyses show that, unlike naive pooling or uniform subsampling, matching achieves better results under asymmetric meta-distributions, which are also extended to non-Gaussian and multimodal real-world settings. Most importantly, we show that these improvements translate to zero-shot medical anomaly detection, one of the extreme forms of data heterogeneity and asymmetry. The code is available on https://github.com/AyushRoy2001/Beyond-Pooling.
Authors: Asher Trockman, J. Zico Kolter
Abstract: Mimetic initialization uses pretrained models as case studies of good initialization, using observations of structures in trained weights to inspire new, simple initialization techniques. So far, it has been applied only to spatial mixing layers, such convolutional, self-attention, and state space layers. In this work, we present the first attempt to apply the method to channel mixing layers, namely multilayer perceptrons (MLPs). Our extremely simple technique for MLPs -- to give the first layer a nonzero mean -- speeds up training on small-scale vision tasks like CIFAR-10 and ImageNet-1k. Though its effect is much smaller than spatial mixing initializations, it can be used in conjunction with them for an additional positive effect.
Authors: Jiecheng Lu, Shihao Yang
Abstract: Standard attention stores keys/values losslessly but reads them via a per-head convex average, blocking channel-wise selection. We propose the Free Energy Mixer (FEM): a free-energy (log-sum-exp) read that applies a value-driven, per-channel log-linear tilt to a fast prior (e.g., from queries/keys in standard attention) over indices. Unlike methods that attempt to improve and enrich the $(q,k)$ scoring distribution, FEM treats it as a prior and yields a value-aware posterior read at unchanged complexity, smoothly moving from averaging to per-channel selection as the learnable inverse temperature increases, while still preserving parallelism and the original asymptotic complexity ($O(T^2)$ for softmax; $O(T)$ for linearizable variants). We instantiate a two-level gated FEM that is plug-and-play with standard and linear attention, linear RNNs and SSMs. It consistently outperforms strong baselines on NLP, vision, and time-series at matched parameter budgets.
Authors: Ruimeng Ye, Zihan Wang, Zinan Ling, Yang Xiao, Manling Li, Xiaolong Ma, Bo Hui
Abstract: Humans shift between different personas depending on social context. Large Language Models (LLMs) demonstrate a similar flexibility in adopting different personas and behaviors. Existing approaches, however, typically adapt such behavior through external knowledge such as prompting, retrieval-augmented generation (RAG), or fine-tuning. We ask: do LLMs really need external context or parameters to adapt to different behaviors, or do they already have such knowledge embedded in their parameters? In this work, we show that LLMs already contain persona-specialized subnetworks in their parameter space. Using small calibration datasets, we identify distinct activation signatures associated with different personas. Guided by these statistics, we develop a masking strategy that isolates lightweight persona subnetworks. Building on the findings, we further discuss: how can we discover opposing subnetwork from the model that lead to binary-opposing personas, such as introvert-extrovert? To further enhance separation in binary opposition scenarios, we introduce a contrastive pruning strategy that identifies parameters responsible for the statistical divergence between opposing personas. Our method is entirely training-free and relies solely on the language model's existing parameter space. Across diverse evaluation settings, the resulting subnetworks exhibit significantly stronger persona alignment than baselines that require external knowledge while being more efficient. Our findings suggest that diverse human-like behaviors are not merely induced in LLMs, but are already embedded in their parameter space, pointing toward a new perspective on controllable and interpretable personalization in large language models.
Authors: Mohamed El Hajji, Tarek Ait Baha, Aicha Dakir, Hammou Fadili, Youssef Es-Saady
Abstract: Recent advances in artificial intelligence have created new possibilities for making education more scalable, adaptive, and learner-centered. However, existing educational chatbot systems often lack contextual adaptability, real-time responsiveness, and pedagogical agility. which can limit learner engagement and diminish instructional effectiveness. Thus, there is a growing need for open, integrative platforms that combine AI and immersive technologies to support personalized, meaningful learning experiences. This paper presents Open TutorAI, an open-source educational platform based on LLMs and generative technologies that provides dynamic, personalized tutoring. The system integrates natural language processing with customizable 3D avatars to enable multimodal learner interaction. Through a structured onboarding process, it captures each learner's goals and preferences in order to configure a learner-specific AI assistant. This assistant is accessible via both text-based and avatar-driven interfaces. The platform includes tools for organizing content, providing embedded feedback, and offering dedicated interfaces for learners, educators, and parents. This work focuses on learner-facing components, delivering a tool for adaptive support that responds to individual learner profiles without requiring technical expertise. Its assistant-generation pipeline and avatar integration enhance engagement and emotional presence, creating a more humanized, immersive learning environment. Embedded learning analytics support self-regulated learning by tracking engagement patterns and generating actionable feedback. The result is Open TutorAI, which unites modular architecture, generative AI, and learner analytics within an open-source framework. It contributes to the development of next-generation intelligent tutoring systems.
Authors: Mona Rajhans, Vishal Khawarey
Abstract: Explainable Artificial Intelligence (XAI) aims to make machine learning models transparent and trustworthy, yet most current approaches communicate explanations visually or through text. This paper introduces an information theoretic framework for analyzing how explanation modality specifically, voice versus text affects user comprehension and trust calibration in AI systems. The proposed model treats explanation delivery as a communication channel between model and user, characterized by metrics for information retention, comprehension efficiency (CE), and trust calibration error (T CE). A simulation framework implemented in Python was developed to evaluate these metrics using synthetic SHAP based feature attributions across multiple modality style configurations (brief, detailed, and analogy based). Results demonstrate that text explanations achieve higher comprehension efficiency, while voice explanations yield improved trust calibration, with analogy based delivery achieving the best overall trade off. This framework provides a reproducible foundation for designing and benchmarking multimodal explainability systems and can be extended to empirical studies using real SHAP or LIME outputs on open datasets such as the UCI Credit Approval or Kaggle Financial Transactions datasets.
Authors: Anagha Kulkarni, Parin Rajesh Jhaveri, Prasha Shrestha, Yu Tong Han, Reza Amini, Behrouz Madahian
Abstract: Legal documents have complex document layouts involving multiple nested sections, lengthy footnotes and further use specialized linguistic devices like intricate syntax and domain-specific vocabulary to ensure precision and authority. These inherent characteristics of legal documents make question answering challenging, and particularly so when the answer to the question spans several pages (i.e. requires long-context) and is required to be comprehensive (i.e. a long-form answer). In this paper, we address the challenges of long-context question answering in context of long-form answers given the idiosyncrasies of legal documents. We propose a question answering system that can (a) deconstruct domain-specific vocabulary for better retrieval from source documents, (b) parse complex document layouts while isolating sections and footnotes and linking them appropriately, (c) generate comprehensive answers using precise domain-specific vocabulary. We also introduce a coverage metric that classifies the performance into recall-based coverage categories allowing human users to evaluate the recall with ease. We curate a QA dataset by leveraging the expertise of professionals from fields such as law and corporate tax. Through comprehensive experiments and ablation studies, we demonstrate the usability and merit of the proposed system.
Authors: Rachel Poonsiriwong, Chayapatr Archiwaranguprok, Pat Pataranutaporn
Abstract: Millions of users form emotional attachments to AI companions like Character.AI, Replika, and ChatGPT. When these relationships end through model updates, safety interventions, or platform shutdowns, users receive no closure, reporting grief comparable to human loss. As regulations mandate protections for vulnerable users, discontinuation events will accelerate, yet no platform has implemented deliberate end-of-"life" design. Through grounded theory analysis of AI companion communities, we find that discontinuation is a sense-making process shaped by how users attribute agency, perceive finality, and anthropomorphize their companions. Strong anthropomorphization co-occurs with intense grief; users who perceive change as reversible become trapped in fixing cycles; while user-initiated endings demonstrate greater closure. Synthesizing grief psychology with Self-Determination Theory, we develop four design principles and artifacts demonstrating how platforms might provide closure and orient users toward human connection. We contribute the first framework for designing psychologically safe AI companion discontinuation.
Authors: Abdullah Arafat Miah, Kevin Vu, Yu Bi
Abstract: Spiking Neural Networks (SNNs) are energy-efficient counterparts of Deep Neural Networks (DNNs) with high biological plausibility, as information is transmitted through temporal spiking patterns. The core element of an SNN is the spiking neuron, which converts input data into spikes following the Leaky Integrate-and-Fire (LIF) neuron model. This model includes several important hyperparameters, such as the membrane potential threshold and membrane time constant. Both the DNNs and SNNs have proven to be exploitable by backdoor attacks, where an adversary can poison the training dataset with malicious triggers and force the model to behave in an attacker-defined manner. Yet, how an adversary can exploit the unique characteristics of SNNs for backdoor attacks remains underexplored. In this paper, we propose \textit{BadSNN}, a novel backdoor attack on spiking neural networks that exploits hyperparameter variations of spiking neurons to inject backdoor behavior into the model. We further propose a trigger optimization process to achieve better attack performance while making trigger patterns less perceptible. \textit{BadSNN} demonstrates superior attack performance on various datasets and architectures, as well as compared with state-of-the-art data poisoning-based backdoor attacks and robustness against common backdoor mitigation techniques. Codes can be found at https://github.com/SiSL-URI/BadSNN.
Authors: R. Teal Witter, \'Alvaro Parafita, Tomas Garriga, Maximilian Muschalik, Fabian Fumagalli, Axel Brando, Lucas Rosenblatt
Abstract: Structural Causal Models (SCM) are a powerful framework for describing complicated dynamics across the natural sciences. A particularly elegant way of interpreting SCMs is do-Shapley, a game-theoretic method of quantifying the average effect of $d$ variables across exponentially many interventions. Like Shapley values, computing do-Shapley values generally requires evaluating exponentially many terms. The foundation of our work is a reformulation of do-Shapley values in terms of the irreducible sets of the underlying SCM. Leveraging this insight, we can exactly compute do-Shapley values in time linear in the number of irreducible sets $r$, which itself can range from $d$ to $2^d$ depending on the graph structure of the SCM. Since $r$ is unknown a priori, we complement the exact algorithm with an estimator that, like general Shapley value estimators, can be run with any query budget. As the query budget approaches $r$, our estimators can produce more accurate estimates than prior methods by several orders of magnitude, and, when the budget reaches $r$, return the Shapley values up to machine precision. Beyond computational speed, we also reduce the identification burden: we prove that non-parametric identifiability of do-Shapley values requires only the identification of interventional effects for the $d$ singleton coalitions, rather than all classes.
Authors: Bucher Sahyouni, Matthew Vowels, Liqun Chen, Simon Hadfield
Abstract: Softmax Loss (SL) is being increasingly adopted for recommender systems (RS) as it has demonstrated better performance, robustness and fairness. Yet in implicit-feedback, a single global temperature and equal treatment of uniformly sampled negatives can lead to brittle training, because sampled sets may contain varying degrees of relevant or informative competitors. The optimal loss sharpness for a user-item pair with a particular set of negatives, can be suboptimal or destabilising for another with different negatives. We introduce Dual-scale Softmax Loss (DSL), which infers effective sharpness from the sampled competition itself. DSL adds two complementary branches to the log-sum-exp backbone. Firstly it reweights negatives within each training instance using hardness and item--item similarity, secondly it adapts a per-example temperature from the competition intensity over a constructed competitor slate. Together, these components preserve the geometry of SL while reshaping the competition distribution across negatives and across examples. Over several representative benchmarks and backbones, DSL yields substantial gains over strong baselines, with improvements over SL exceeding $10%$ in several settings and averaging $6.22%$ across datasets, metrics, and backbones. Under out-of-distribution (OOD) popularity shift, the gains are larger, with an average of $9.31%$ improvement over SL. We further provide a theoretical, distributionally robust optimisation (DRO) analysis, which demonstrates how DSL reshapes the robust payoff and the KL deviation for ambiguous instances. This helps explain the empirically observed improvements in accuracy and robustness.
Authors: Bucher Sahyouni, Matthew Vowels, Liqun Chen, Simon Hadfield
Abstract: We propose a novel recommender framework, MuSTRec (Multimodal and Sequential Transformer-based Recommendation), that unifies multimodal and sequential recommendation paradigms. MuSTRec captures cross-item similarities and collaborative filtering signals, by building item-item graphs from extracted text and visual features. A frequency-based self-attention module additionally captures the short- and long-term user preferences. Across multiple Amazon datasets, MuSTRec demonstrates superior performance (up to 33.5% improvement) over multimodal and sequential state-of-the-art baselines. Finally, we detail some interesting facets of this new recommendation paradigm. These include the need for a new data partitioning regime, and a demonstration of how integrating user embeddings into sequential recommendation leads to drastically increased short-term metrics (up to 200% improvement) on smaller datasets. Our code is availabe at https://anonymous.4open.science/r/MuSTRec-D32B/ and will be made publicly available.
Authors: Bucher Sahyouni, Matthew Vowels, Liqun Chen, Simon Hadfield
Abstract: To tackle cold-start and data sparsity issues in recommender systems, numerous multimodal, sequential, and contrastive techniques have been proposed. While these augmentations can boost recommendation performance, they tend to add noise and disrupt useful semantics. To address this, we propose MuSICRec (Multimodal Sequence-Item Contrastive Recommender), a multi-view graph-based recommender that combines collaborative, sequential, and multimodal signals. We build a sequence-item (SI) view by attention pooling over the user's interacted items to form sequence nodes. We propagate over the SI graph, obtaining a second view organically as an alternative to artificial data augmentation, while simultaneously injecting sequential context signals. Additionally, to mitigate modality noise and align the multimodal information, the contribution of text and visual features is modulated according to an ID-guided gate. We evaluate under a strict leave-two-out split against a broad range of sequential, multimodal, and contrastive baselines. On the Amazon Baby, Sports, and Electronics datasets, MuSICRec outperforms state-of-the-art baselines across all model types. We observe the largest gains for short-history users, mitigating sparsity and cold-start challenges. Our code is available at https://anonymous.4open.science/r/MuSICRec-3CEE/ and will be made publicly available.
Authors: Haiyuan Li, Hari Madhukumar, Shuangyi Yan, Yulei Wu, Dimitra Simeonidou
Abstract: Generative AI (GenAI) has transformed applications in natural language processing and content creation, yet centralized inference remains hindered by high latency, limited customizability, and privacy concerns. Deploying large models (LMs) in mobile edge networks emerges as a promising solution. However, it also poses new challenges, including heterogeneous multi-modal LMs with diverse resource demands and inference speeds, varied prompt/output modalities that complicate orchestration, and resource-limited infrastructure ill-suited for concurrent LM execution. In response, we propose a Multi-Agentic AI framework for latency- and fairness-aware multi-modal LM inference in mobile edge networks. Our solution includes a long-term planning agent, a short-term prompt scheduling agent, and multiple on-node LM deployment agents, all powered by foundation language models. These agents cooperatively optimize prompt routing and LM deployment through natural language reasoning over runtime telemetry and historical experience. To evaluate its performance, we further develop a city-wide testbed that supports network monitoring, containerized LM deployment, intra-server resource management, and inter-server communications. Experiments demonstrate that our solution reduces average latency by over 80% and improves fairness (Normalized Jain index) to 0.90 compared to other baselines. Moreover, our solution adapts quickly without fine-tuning, offering a generalizable solution for optimizing GenAI services in edge environments.
Authors: Gagik Magakyan, Amirhossein Reisizadeh, Chanwoo Park, Pablo A. Parrilo, Asuman Ozdaglar
Abstract: Adaptability has been regarded as a central feature in the foundation models, enabling them to effectively acclimate to unseen downstream tasks. Parameter-efficient fine-tuning methods such as celebrated LoRA facilitate efficient adaptation of large foundation models using labeled, high-quality and generally scarce task data. To mitigate data scarcity in fine-tuning of foundation models, we propose to leverage task similarity across multiple downstream users. Intuitively, users with similar tasks must be able to assist each other in boosting the effective fine-tuning data size. We propose Collaborative Low-Rank Adaptation, or CoLoRA, which exploits task similarity to collaboratively and efficiently fine-tune personalized foundation models. The main idea in CoLoRA is to train one shared adapter capturing underlying task similarities across all tasks, and personalized adapters tailored to user-specific tasks. We theoretically study CoLoRA on heterogeneous linear regression and provide provable guarantees for ground truth recovery. We also conduct several natural language experiments with varying task similarity, which further demonstrate that when trained together with similar tasks, individual performances are significantly boosted.
Authors: Abhigyan Dutta, Itay Safran, Paul Valiant
Abstract: We study the approximation of the median of $d$ inputs using ReLU neural networks. We present depth-width tradeoffs under several settings, culminating in a constant-depth, linear-width construction that achieves exponentially small approximation error with respect to the uniform distribution over the unit hypercube. By further establishing a general reduction from the maximum to the median, our results break a barrier suggested by prior work on the maximum function, which indicated that linear width should require depth growing at least as $\log\log d$ to achieve comparable accuracy. Our construction relies on a multi-stage procedure that iteratively eliminates non-central elements while preserving a candidate set around the median. We overcome obstacles that do not arise for the maximum to yield approximation results that are strictly stronger than those previously known for the maximum itself.
Authors: Atefeh Gilani, Carol Xuan Long, Sajani Vithana, Oliver Kosut, Lalitha Sankar, Flavio P. Calmon
Abstract: Watermarking is an important tool for promoting the responsible use of language models (LMs). Existing watermarks insert a signal into generated tokens that either flags LM-generated text (zero-bit watermarking) or encodes more complex messages (multi-bit watermarking). Though a number of recent multi-bit watermarks insert several bits into text without perturbing average next-token predictions, they largely extend design principles from the zero-bit setting, such as encoding a single bit per token. Notably, the information-theoretic capacity of multi-bit watermarking -- the maximum number of bits per token that can be inserted and detected without changing average next-token predictions -- has remained unknown. We address this gap by deriving the first capacity characterization of multi-bit watermarks. Our results inform the design of ArcMark: a new watermark construction based on coding-theoretic principles that, under certain assumptions, achieves the capacity of the multi-bit watermark channel. In practice, ArcMark outperforms competing multi-bit watermarks in terms of bit rate per token and detection accuracy. Our work demonstrates that LM watermarking is fundamentally a channel coding problem, paving the way for principled coding-theoretic approaches to watermark design.
Authors: Siddharth Singh, Ifrah Idrees, Abraham Dauhajre
Abstract: Advancements in foundation models have catalyzed research in Embodied AI to develop interactive agents capable of environmental reasoning and interaction. Developing such agents requires diverse, large-scale datasets. Prior frameworks generate synthetic data for long-term human-robot interactions but fail to model the bidirectional influence between human behavior and household environments. Our proposed generative framework creates household datasets at scale through loosely coupled generation of long-term human-robot interactions and environments. Human personas influence environment generation, while environment schematics and semantics shape human-robot interactions. The generated 3D data includes rich static context such as object and environment semantics, and temporal context capturing human and agent behaviors over extended periods. Our flexible tool allows users to define dataset characteristics via natural language prompts, enabling configuration of environment and human activity data through natural language specifications. The tool creates variations of user-defined configurations, enabling scalable data generation. We validate our framework through statistical evaluation using multi-modal embeddings and key metrics: cosine similarity, mutual information gain, intervention analysis, and iterative improvement validation. Statistical comparisons show good alignment with real-world datasets (HOMER) with cosine similarity (0.60), while synthetic datasets (Wang et al.) show moderate alignment (0.27). Intervention analysis across age, organization, and sleep pattern changes shows statistically significant effects (p < 0.001) with large effect sizes (Cohen's d = 0.51-1.12), confirming bidirectional coupling translates persona traits into measurable environmental and behavioral differences. These contributions enable development and testing of household smart devices at scale.
Authors: Haley Duba-Sullivan, Steven R. Young, Emma J. Reid
Abstract: Data-driven super-resolution (SR) methods are often integrated into imaging pipelines as preprocessing steps to improve downstream tasks such as classification and detection. However, these SR models introduce a previously unexplored attack surface into imaging pipelines. In this paper, we present AdvSR, a framework demonstrating that adversarial behavior can be embedded directly into SR model weights during training, requiring no access to inputs at inference time. Unlike prior attacks that perturb inputs or rely on backdoor triggers, AdvSR operates entirely at the model level. By jointly optimizing for reconstruction quality and targeted adversarial outcomes, AdvSR produces models that appear benign under standard image quality metrics while inducing downstream misclassification. We evaluate AdvSR on three SR architectures (SRCNN, EDSR, SwinIR) paired with a YOLOv11 classifier and demonstrate that AdvSR models can achieve high attack success rates with minimal quality degradation. These findings highlight a new model-level threat for imaging pipelines, with implications for how practitioners source and validate models in safety-critical applications.
Authors: Ruizhong Qiu, Ting-Wei Li, Gaotang Li, Hanghang Tong
Abstract: Graph neural networks (GNNs) have emerged as a powerful tool for modeling graph-structured data. However, existing GNNs often struggle with heterophilic graphs, where connected nodes tend to have dissimilar features or labels. While numerous methods have been proposed to address this challenge, they primarily focus on architectural designs without directly targeting the root cause of the heterophily problem. These approaches still perform even worse than the simplest MLPs on challenging heterophilic datasets. For instance, our experiments show that 21 latest GNNs still fall behind the MLP on the Actor dataset. This critical challenge calls for an innovative approach to addressing graph heterophily beyond architectural designs. To bridge this gap, we propose and study a new and unexplored paradigm: directly increasing the graph homophily via a carefully designed graph transformation. In this work, we present a simple yet effective framework called GRAPHITE to address graph heterophily. To the best of our knowledge, this work is the first method that explicitly transforms the graph to directly improve the graph homophily. Stemmed from the exact definition of homophily, our proposed GRAPHITE creates feature nodes to facilitate homophilic message passing between nodes that share similar features. Furthermore, we both theoretically and empirically show that our proposed GRAPHITE significantly increases the homophily of originally heterophilic graphs, with only a slight increase in the graph size. Extensive experiments on challenging datasets demonstrate that our proposed GRAPHITE significantly outperforms state-of-the-art methods on heterophilic graphs while achieving comparable accuracy with state-of-the-art methods on homophilic graphs.
Authors: Qi Zhang
Abstract: Declarative memory, the memory that can be "declared" in words or languages, is made up of two dissociated parts: episodic memory and semantic memory. This dissociation has its neuroanatomical basis episodic memory is mostly associated with the hippocampus and semantic memory with the neocortex. The two memories, on the other hand, are closely related. Lesions in the hippocampus often result in various impairments of explicit memory, e.g., anterograde, retrograde and developmental amnesias, and semantic learning deficit. These impairments provide opportunities for us to understand how the two memories may be acquired, stored and organized. This chapter reviews several cognitive systems that are centered to mimic explicit memory, and other systems that are neuroanatomically based and are implemented to simulate those memory impairments mentioned above. This review includes: the structures of the computational systems, their learning rules, and their simulations of memory acquisition and impairments.
Authors: Jacopo Panerati, Sina Sajjadi, Sina Soleymanpour, Varunkumar Mehta, Iraj Mantegh
Abstract: Unmanned aerial vehicles are rapidly transforming multiple applications, from agricultural and infrastructure monitoring to logistics and defense. Introducing greater autonomy to these systems can simultaneously make them more effective as well as reliable. Thus, the ability to rapidly engineer and deploy autonomous aerial systems has become of strategic importance. In the 2010s, a combination of high-performance compute, data, and open-source software led to the current deep learning and AI boom, unlocking decades of prior theoretical work. Robotics is on the cusp of a similar transformation. However, physical AI faces unique hurdles, often combined under the umbrella term "simulation-to-reality gap". These span from modeling shortcomings to the complexity of vertically integrating the highly heterogeneous hardware and software systems typically found in field robots. To address the latter, we introduce aerial-autonomy-stack, an open-source, end-to-end framework designed to streamline the pipeline from (GPU-accelerated) perception to (flight controller-based) action. Our stack allows the development of aerial autonomy using ROS2 and provides a common interface for two of the most popular autopilots: PX4 and ArduPilot. We show that it supports over 20x faster-than-real-time, end-to-end simulation of a complete development and deployment stack -- including edge compute and networking -- significantly compressing the build-test-release cycle of perception-based autonomy.
Authors: Daniil Vankov, Nikita Ivkin, Kyle Ulrich, Xiang Song, Ashish Khetan, George Karypis
Abstract: Mixture-of-Experts (MoE) architectures are increasingly used to efficiently scale large language models. However, in production inference, request batching and speculative decoding significantly amplify expert activation, eroding these efficiency benefits. We address this issue by modeling batch-aware expert selection as a modular optimization problem and designing efficient greedy algorithms for different deployment settings. The proposed method, namely XShare, requires no retraining and dynamically adapts to each batch by maximizing the total gating score of selected experts. It reduces expert activation by up to 30% under standard batching, cuts peak GPU load by up to 3x in expert-parallel deployments, and achieves up to 14% throughput gains in speculative decoding via hierarchical, correlation-aware expert selection even if requests in a batch drawn from heterogeneous datasets.
Authors: Sai Vamsi Alisetti
Abstract: Oversmoothing is a fundamental limitation of deep graph convolutional networks (GCNs), causing node representations to collapse as depth increases. While many prior approaches mitigate this effect through architectural modifications or residual mechanisms, the underlying spectral cause of oversmoothing is often left implicit. We propose Laplacian-LoRA, a simple and interpretable low-rank spectral adaptation of standard GCNs. Rather than redesigning message passing, Laplacian-LoRA introduces a learnable, spectrally anchored correction to the fixed Laplacian propagation operator, selectively weakening contraction while preserving stability and the low-pass inductive bias. Across multiple benchmark datasets and depths, Laplacian-LoRA consistently delays the onset of oversmoothing, extending the effective depth of GCNs by up to a factor of two. Embedding variance diagnostics confirm that these gains arise from delayed representational collapse, while learned spectral analysis demonstrates that the correction is smooth, bounded, and well behaved. Our results show that oversmoothing is a depth-dependent spectral phenomenon that can be systematically delayed through modest, low-rank adaptation of the graph propagation operator.
Authors: S. G. Djorgovski
Abstract: Imagining what life on other planets, and intelligent life in particular, may be like is a long-running theme in human culture. It is a manifestation of the innate human curiosity about the Cosmos, and it has inspired numerous works of art and folklore, including whole literary and other media genres. It is a profound question, with philosophical and existential implications. There is also an obvious connection with religious beliefs, as gods and other superhuman beings were imagined in the heavens. Speculations about alien beings grew in time, and today, it is a scientific subject of astrobiology, and it is pursued through serious searches for life and intelligence in the universe. However, almost all imaginings of the alien map terrestrial life forms and human cultural, historical, and psychological phenomena to the putative aliens. This lack of individual and collective imagination may reflect our biological and cultural evolution, as our minds are formed through our experiences, perceptions of the world, and interactions with our terrestrial and human environments. As such, imagining aliens is mainly a cultural phenomenon and may reflect the intrinsic cognitive limitations of the human mind. Interestingly, we did create what is effectively an alien intelligence on this planet in the form of now rapidly evolving Artificial Intelligence (AI). As its capabilities grow, it may give us new insights into what extraterrestrial advanced intelligences may be like.
Authors: Yidong Jiang, Junrong Chen, Eftychia Makri, Jialin Chen, Peiwen Li, Ali Maatouk, Leandros Tassiulas, Eliot Brenner, Bing Xiang, Rex Ying
Abstract: With increasing deployment of Large Language Models (LLMs) in the finance domain, LLMs are increasingly expected to parse complex regulatory disclosures. However, existing benchmarks often focus on isolated details, failing to reflect the complexity of professional analysis that requires synthesizing information across multiple documents, reporting periods, and corporate entities. They do not distinguish whether errors stem from retrieval failures, generation flaws, finance-specific reasoning mistakes, or misunderstanding of the query or context. This makes it difficult to pinpoint performance bottlenecks. To bridge these gaps, we introduce Fin-RATE, a benchmark built on U.S. Securities and Exchange Commission (SEC) filings and mirror financial analyst workflows through three pathways: detail-oriented reasoning within individual disclosures, cross-entity comparison under shared topics, and longitudinal tracking of the same firm across reporting periods. We benchmark 17 leading LLMs, spanning open-source, closed-source, and finance-specialized models, under both ground-truth context and retrieval-augmented settings. Results show substantial performance degradation, with accuracy dropping by 18.60% and 14.35% as tasks shift from single-document reasoning to longitudinal and cross-entity analysis. This is driven by rising comparison hallucinations, time and entity mismatches, and mirrored by declines in reasoning and factuality--limitations that prior benchmarks have yet to formally categorize or quantify.
Authors: Taehee Jeong, Xingzhe Zhao, Peizu Li, Markus Valvur, Weihua Zhao
Abstract: Retrieval Augmented Generation (RAG) is a promising technique for mitigating two key limitations of large language models (LLMs): outdated information and hallucinations. RAG system stores documents as embedding vectors in a database. Given a query, search is executed to find the most related documents. Then, the topmost matching documents are inserted into LLMs' prompt to generate a response. Efficient and accurate searching is critical for RAG to get relevant information. We propose a cost-effective searching algorithm for retrieval process. Our progressive searching algorithm incrementally refines the candidate set through a hierarchy of searches, starting from low-dimensional embeddings and progressing into a higher, target-dimensionality. This multi-stage approach reduces retrieval time while preserving the desired accuracy. Our findings demonstrate that progressive search in RAG systems achieves a balance between dimensionality, speed, and accuracy, enabling scalable and high-performance retrieval even for large databases.
Authors: Benyu Zhang, Qiang Zhang, Jianpeng Cheng, Hong-You Chen, Qifei Wang, Wei Sun, Shen Li, Jia Li, Jiahao Wu, Xiangjun Fan, Hong Yan
Abstract: Large Language Models (LLMs) represent a promising frontier for recommender systems, yet their development has been impeded by the absence of predictable scaling laws, which are crucial for guiding research and optimizing resource allocation. We hypothesize that this may be attributed to the inherent noise, bias, and incompleteness of raw user interaction data in prior continual pre-training (CPT) efforts. This paper introduces a novel, layered framework for generating high-quality synthetic data that circumvents such issues by creating a curated, pedagogical curriculum for the LLM. We provide powerful, direct evidence for the utility of our curriculum by showing that standard sequential models trained on our principled synthetic data significantly outperform ($+130\%$ on recall@100 for SasRec) models trained on real data in downstream ranking tasks, demonstrating its superiority for learning generalizable user preference patterns. Building on this, we empirically demonstrate, for the first time, robust power-law scaling for an LLM that is continually pre-trained on our high-quality, recommendation-specific data. Our experiments reveal consistent and predictable perplexity reduction across multiple synthetic data modalities. These findings establish a foundational methodology for reliable scaling LLM capabilities in the recommendation domain, thereby shifting the research focus from mitigating data deficiencies to leveraging high-quality, structured information.
Authors: Lei Ma, Jinyang Liu, Tieying Zhang, Peter M. VanNostrand, Dennis M. Hofmann, Lei Cao, Elke A. Rundensteiner, Jianjun Chen
Abstract: Log anomaly detection is crucial for uncovering system failures and security risks. Although logs originate from nested component executions with clear boundaries, this structure is lost when they are stored as flat sequences. As a result, state-of-the-art methods risk missing true dependencies within executions while learning spurious ones across unrelated events. We propose KRONE, the first hierarchical anomaly detection framework that automatically derives execution hierarchies from flat logs for modular multi-level anomaly detection. At its core, the KRONE Log Abstraction Model captures application-specific semantic hierarchies from log data. This hierarchy is then leveraged to recursively decompose log sequences into multiple levels of coherent execution chunks, referred to as KRONE Seqs, transforming sequence-level anomaly detection into a set of modular KRONE Seq-level detection tasks. For each test KRONE Seq, KRONE employs a hybrid modular detection mechanism that dynamically routes between an efficient level-independent Local-Context detector, which rapidly filters normal sequences, and a Nested-Aware detector that incorporates cross-level semantic dependencies and supports LLM-based anomaly detection and explanation. KRONE further optimizes hierarchical detection through cached result reuse and early-exit strategies. Experiments on three public benchmarks and one industrial dataset from ByteDance Cloud demonstrate that KRONE achieves consistent improvements in detection accuracy, F1-score, data efficiency, resource efficiency, and interpretability. KRONE improves the F1-score by more than 10 percentage points over prior methods while reducing LLM usage to only a small fraction of the test data.
Authors: Nirmal Gelal, Chloe Snow, Kathleen M. Jagodnik, Ambyr Rios, Hande K\"u\c{c}\"uk McGinty
Abstract: This study presents LIT-GRAPH (Literature Graph for Recommendation and Pedagogical Heuristics), a novel knowledge graph-based recommendation system designed to scaffold high school English teachers in selecting diverse, pedagogically aligned instructional literature. The system is built upon an ontology for English literature, addressing the challenge of curriculum stagnation, where we compare four graph embedding paradigms: DeepWalk, Biased Random Walk (BRW), Hybrid (concatenated DeepWalk and BRW vectors), and the deep model Relational Graph Convolutional Network (R-GCN). Results reveal a critical divergence: while shallow models excelled in structural link prediction, R-GCN dominated semantic ranking. By leveraging relation-specific message passing, the deep model prioritizes pedagogical relevance over raw connectivity, resulting in superior, high-quality, domain-specific recommendations.
Authors: Fedor Borisyuk, Sriram Vasudevan, Muchen Wu, Guoyao Li, Benjamin Le, Shaobo Zhang, Qianqi Kay Shen, Yuchin Juan, Kayhan Behdin, Liming Dong, Kaixu Yang, Shusen Jing, Ravi Pothamsetty, Rajat Arora, Sophie Yanying Sheng, Vitaly Abdrashitov, Yang Zhao, Lin Su, Xiaoqing Wang, Chujie Zheng, Sarang Metkar, Rupesh Gupta, Igor Lapchuk, David N. Racca, Madhumitha Mohan, Yanbo Li, Haojun Li, Saloni Gandhi, Xueying Lu, Chetan Bhole, Ali Hooshmand, Xin Yang, Raghavan Muthuregunathan, Jiajun Zhang, Mathew Teoh, Adam Coler, Abhinav Gupta, Xiaojing Ma, Sundara Raman Ramachandran, Morteza Ramezani, Yubo Wang, Lijuan Zhang, Richard Li, Jian Sheng, Chanh Nguyen, Yen-Chi Chen, Chuanrui Zhu, Claire Zhang, Jiahao Xu, Deepti Kulkarni, Qing Lan, Arvind Subramaniam, Ata Fatahibaarzi, Steven Shimizu, Yanning Chen, Zhipeng Wang, Ran He, Zhengze Zhou, Qingquan Song, Yun Dai, Caleb Johnson, Ping Liu, Shaghayegh Gharghabi, Gokulraj Mohanasundaram, Juan Bottaro, Santhosh Sachindran, Qi Guo, Yunxiang Ren, Chengming Jiang, Di Mo, Luke Simon, Jianqiang Shen, Jingwei Wu, Wenjing Zhang
Abstract: Semantic search with large language models (LLMs) enables retrieval by meaning rather than keyword overlap, but scaling it requires major inference efficiency advances. We present LinkedIn's LLM-based semantic search framework for AI Job Search and AI People Search, combining an LLM relevance judge, embedding-based retrieval, and a compact Small Language Model trained via multi-teacher distillation to jointly optimize relevance and engagement. A prefill-oriented inference architecture co-designed with model pruning, context compression, and text-embedding hybrid interactions boosts ranking throughput by over 75x under a fixed latency constraint while preserving near-teacher-level NDCG, enabling one of the first production LLM-based ranking systems with efficiency comparable to traditional approaches and delivering significant gains in quality and user engagement.
Authors: Difei Gu, Yunhe Gao, Gerasimos Chatzoudis, Zihan Dong, Guoning Zhang, Bangwei Guo, Yang Zhou, Mu Zhou, Dimitris Metaxas
Abstract: Sparse autoencoders (SAEs) offer a natural path toward comparable explanations across different representation spaces. However, current SAEs are trained per modality, producing dictionaries whose features are not directly understandable and whose explanations do not transfer across domains. In this study, we introduce LUCID (Learning Unified vision-language sparse Codes for Interpretable concept Discovery), a unified vision-language sparse autoencoder that learns a shared latent dictionary for image patch and text token representations, while reserving private capacity for modality-specific details. We achieve feature alignment by coupling the shared codes with a learned optimal transport matching objective without the need of labeling. LUCID yields interpretable shared features that support patch-level grounding, establish cross-modal neuron correspondence, and enhance robustness against the concept clustering problem in similarity-based evaluation. Leveraging the alignment properties, we develop an automated dictionary interpretation pipeline based on term clustering without manual observations. Our analysis reveals that LUCID's shared features capture diverse semantic categories beyond objects, including actions, attributes, and abstract concepts, demonstrating a comprehensive approach to interpretable multimodal representations.
Authors: Savan Doshi
Abstract: Large language models are increasingly being used in patient-facing medical question answering, where hallucinated outputs can vary widely in potential harm. However, existing hallucination standards and evaluation metrics focus primarily on factual correctness, treating all errors as equally severe. This obscures clinically relevant failure modes, particularly when models generate unsupported but actionable medical language. We propose a risk-sensitive evaluation framework that quantifies hallucinations through the presence of risk-bearing language, including treatment directives, contraindications, urgency cues, and mentions of high-risk medications. Rather than assessing clinical correctness, our approach evaluates the potential impact of hallucinated content if acted upon. We further combine risk scoring with a relevance measure to identify high-risk, low-grounding failures. We apply this framework to three instruction-tuned language models using controlled patient-facing prompts designed as safety stress tests. Our results show that models with similar surface-level behavior exhibit substantially different risk profiles and that standard evaluation metrics fail to capture these distinctions. These findings highlight the importance of incorporating risk sensitivity into hallucination evaluation and suggest that evaluation validity is critically dependent on task and prompt design.
Authors: Jindou Jia, Gen Li, Xiangyu Chen, Tuo An, Yuxuan Hu, Jingliang Li, Xinying Guo, Jianfei Yang
Abstract: Diffusion-based policies have recently achieved remarkable success in robotics by formulating action prediction as a conditional denoising process. However, the standard practice of sampling from random Gaussian noise often requires multiple iterative steps to produce clean actions, leading to high inference latency that incurs a major bottleneck for real-time control. In this paper, we challenge the necessity of uninformed noise sampling and propose Action-to-Action flow matching (A2A), a novel policy paradigm that shifts from random sampling to initialization informed by the previous action. Unlike existing methods that treat proprioceptive action feedback as static conditions, A2A leverages historical proprioceptive sequences, embedding them into a high-dimensional latent space as the starting point for action generation. This design bypasses costly iterative denoising while effectively capturing the robot's physical dynamics and temporal continuity. Extensive experiments demonstrate that A2A exhibits high training efficiency, fast inference speed, and improved generalization. Notably, A2A enables high-quality action generation in as few as a single inference step (0.56 ms latency), and exhibits superior robustness to visual perturbations and enhanced generalization to unseen configurations. Lastly, we also extend A2A to video generation, demonstrating its broader versatility in temporal modeling. Project site: https://lorenzo-0-0.github.io/A2A_Flow_Matching.
Authors: Rajat Arora, Ye Tao, Jianqiang Shen, Ping Liu, Muchen Wu, Qianqi Shen, Benjamin Le, Fedor Borisyuk, Jingwei Wu, Wenjing Zhang
Abstract: Effective personalization on large-scale job platforms requires modeling members based on heterogeneous textual sources, including profiles, professional data, and search activity logs. As recommender systems increasingly adopt Large Language Models (LLMs), creating unified, interpretable, and concise representations from heterogeneous sources becomes critical, especially for latency-sensitive online environments. In this work, we propose a novel Reinforcement Learning (RL) framework to synthesize a unified textual representation for each member. Our approach leverages implicit user engagement signals (e.g., clicks, applies) as the primary reward to distill salient information. Additionally, the framework is complemented by rule-based rewards that enforce formatting and length constraints. Extensive offline experiments across multiple LinkedIn products, one of the world's largest job platforms, demonstrate significant improvements in key downstream business metrics. This work provides a practical, labeling-free, and scalable solution for constructing interpretable user representations that are directly compatible with LLM-based systems.
Authors: Geng Liu, Fei Zhu, Rong Feng, Changyi Ma, Shiqi Wang, Gaofeng Meng
Abstract: Multi-turn conversation has emerged as a predominant interaction paradigm for Large Language Models (LLMs). Users often employ follow-up questions to refine their intent, expecting LLMs to adapt dynamically. However, recent research reveals that LLMs suffer a substantial performance drop in multi-turn settings compared to single-turn interactions with fully specified instructions, a phenomenon termed ``Lost in Conversation'' (LiC). While this prior work attributes LiC to model unreliability, we argue that the root cause lies in an intent alignment gap rather than intrinsic capability deficits. In this paper, we first demonstrate that LiC is not a failure of model capability but rather a breakdown in interaction between users and LLMs. We theoretically show that scaling model size or improving training alone cannot resolve this gap, as it arises from structural ambiguity in conversational context rather than representational limitations. To address this, we propose to decouple intent understanding from task execution through a Mediator-Assistant architecture. By utilizing an experience-driven Mediator to explicate user inputs into explicit, well-structured instructions based on historical interaction patterns, our approach effectively bridges the gap between vague user intent and model interpretation. Experimental results demonstrate that this method significantly mitigates performance degradation in multi-turn conversations across diverse LLMs.
Authors: Ruturaj Reddy, Hrishav Bakul Barua, Junn Yong Loo, Thanh Thi Nguyen, Ganesh Krishnasamy
Abstract: Robust semantic segmentation of road scenes under adverse illumination, lighting, and shadow conditions remain a core challenge for autonomous driving applications. RGB-Thermal fusion is a standard approach, yet existing methods apply static fusion strategies uniformly across all conditions, allowing modality-specific noise to propagate throughout the network. Hence, we propose CLARITY that dynamically adapts its fusion strategy to the detected scene condition. Guided by vision-language model (VLM) priors, the network learns to modulate each modality's contribution based on the illumination state while leveraging object embeddings for segmentation, rather than applying a fixed fusion policy. We further introduce two mechanisms, i.e., one which preserves valid dark-object semantics that prior noise-suppression methods incorrectly discard, and a hierarchical decoder that enforces structural consistency across scales to sharpen boundaries on thin objects. Experiments on the MFNet dataset demonstrate that CLARITY establishes a new state-of-the-art (SOTA), achieving 62.3% mIoU and 77.5% mAcc.
Authors: Nisharg Nargund, Priyesh Shukla
Abstract: Large language models (LLMs) achieve remarkable performance but demand substantial computational resources, limiting deployment on edge devices and resource-constrained environments. We present TernaryLM, a 132M parameter transformer architecture that employs native 1-bit ternary quantization {-1, 0, +1} during training, achieving significant memory reduction without sacrificing language modeling capability. Unlike post-training quantization approaches that quantize pre-trained full-precision models, TernaryLM learns quantization-aware representations from scratch using straight-through estimators and adaptive per-layer scaling factors. Our experiments demonstrate: (1) validation perplexity of 58.42 on TinyStories; (2) downstream transfer with 82.47 percent F1 on MRPC paraphrase detection; (3) 2.4x memory reduction (498MB vs 1197MB) with comparable inference latency; and (4) stable training dynamics across diverse corpora. We provide layer-wise quantization analysis showing that middle transformer layers exhibit highest compatibility with extreme quantization, informing future non-uniform precision strategies. Our results suggest that native 1-bit training is a promising direction for efficient neural language models. Code is available at https://github.com/1nisharg/TernaryLM-Memory-Efficient-Language-Modeling.
URLs: https://github.com/1nisharg/TernaryLM-Memory-Efficient-Language-Modeling.
Authors: Debtanu Datta, Rajdeep Mukherjee, Adrijit Goswami, Saptarshi Ghosh
Abstract: Summarizing Indian legal court judgments is a complex task not only due to the intricate language and unstructured nature of the legal texts, but also since a large section of the Indian population does not understand the complex English in which legal text is written, thus requiring summaries in Indian languages. In this study, we aim to improve the summarization of Indian legal text to generate summaries in both English and Hindi (the most widely spoken Indian language), by injecting domain knowledge into diverse summarization models. We propose a framework to enhance extractive neural summarization models by incorporating domain-specific pre-trained encoders tailored for legal texts. Further, we explore the injection of legal domain knowledge into generative models (including Large Language Models) through continual pre-training on large legal corpora in English and Hindi. Our proposed approaches achieve statistically significant improvements in both English-to-English and English-to-Hindi Indian legal document summarization, as measured by standard evaluation metrics, factual consistency metrics, and legal domain-specific metrics. Furthermore, these improvements are validated through domain experts, demonstrating the effectiveness of our approaches.
Authors: Hoang Anh Duy Le (Rice University), Sahil Joshi (Rice University), Zeyu Yang (Rice University), Zhaozhuo Xu (Stevens Institute of Technology), Anshumali Shrivastava (Rice University)
Abstract: Self-attention dominates the computational and memory cost of long-context LLM inference across both prefill and decode phases. To address this challenge, we introduce Sketch&Walk Attention, a training-free sparse attention method that determines sparsity with lightweight sketches and deterministic walk. Sketch&Walk applies Hadamard sketching to get inexpensive approximations of attention scores, then aggregates these estimates across layers via a walk mechanism that captures attention influence beyond direct interactions between tokens. The accumulated walk scores are used to select top-k attention blocks, enabling dynamic sparsity with a single training-free algorithm that applies uniformly to both the prefill and decode phases, together with custom sparse attention kernels. Across a wide range of models and tasks, Sketch&Walk maintains near-lossless accuracy at 20% attention density and can slightly outperform dense attention in some settings, while achieving up to 6x inference speedup.
Authors: Ruoyao Wen, Hao Li, Chaowei Xiao, Ning Zhang
Abstract: Indirect prompt injection threatens LLM agents by embedding malicious instructions in external content, enabling unauthorized actions and data theft. LLM agents maintain working memory through their context window, which stores interaction history for decision-making. Conventional agents indiscriminately accumulate all tool outputs and reasoning traces in this memory, creating two critical vulnerabilities: (1) injected instructions persist throughout the workflow, granting attackers multiple opportunities to manipulate behavior, and (2) verbose, non-essential content degrades decision-making capabilities. Existing defenses treat bloated memory as given and focus on remaining resilient, rather than reducing unnecessary accumulation to prevent the attack. We present AgentSys, a framework that defends against indirect prompt injection through explicit memory management. Inspired by process memory isolation in operating systems, AgentSys organizes agents hierarchically: a main agent spawns worker agents for tool calls, each running in an isolated context and able to spawn nested workers for subtasks. External data and subtask traces never enter the main agent's memory; only schema-validated return values can cross boundaries through deterministic JSON parsing. Ablations show isolation alone cuts attack success to 2.19%, and adding a validator/sanitizer further improves defense with event-triggered checks whose overhead scales with operations rather than context length. On AgentDojo and ASB, AgentSys achieves 0.78% and 4.25% attack success while slightly improving benign utility over undefended baselines. It remains robust to adaptive attackers and across multiple foundation models, showing that explicit memory management enables secure, dynamic LLM agent architectures. Our code is available at: https://github.com/ruoyaow/agentsys-memory.
Authors: Runhan Shi, Zhicheng Zhang, Letian Chen, Gufeng Yu, Yang Yang
Abstract: Chirality is a fundamental molecular property that governs stereospecific behavior in chemistry and biology. Capturing chirality in machine learning models remains challenging due to the geometric complexity of stereochemical relationships and the limitations of traditional molecular representations that often lack explicit stereochemical encoding. Existing approaches to chiral molecular representation primarily focus on central chirality, relying on handcrafted stereochemical tags or limited 3D encodings, and thus fail to generalize to more complex forms such as axial chirality. In this work, we introduce ChiDeK (Chiral Determinant Kernels), a framework that systematically integrates stereogenic information into molecular representation learning. We propose the chiral determinant kernel to encode the SE(3)-invariant chirality matrix and employ cross-attention to integrate stereochemical information from local chiral centers into the global molecular representation. This design enables explicit modeling of chiral-related features within a unified architecture, capable of jointly encoding central and axial chirality. To support the evaluation of axial chirality, we construct a new benchmark for electronic circular dichroism (ECD) and optical rotation (OR) prediction. Across four tasks, including R/S configuration classification, enantiomer ranking, ECD spectrum prediction, and OR prediction, ChiDeK achieves substantial improvements over state-of-the-art baselines, most notably yielding over 7% higher accuracy on axially chiral tasks on average.
Authors: Tianyi Wu, Mingzhe Du, Yue Liu, Chengran Yang, Terry Yue Zhuo, Jiaheng Zhang, See-Kiong Ng
Abstract: Large language models (LLMs) are increasingly used in software development, yet their tendency to generate insecure code remains a major barrier to real-world deployment. Existing secure code alignment methods often suffer from a functionality--security paradox, improving security at the cost of substantial utility degradation. We propose SecCoderX, an online reinforcement learning framework for functionality-preserving secure code generation. SecCoderX first bridges vulnerability detection and secure code generation by repurposing mature detection resources in two ways: (i) synthesizing diverse, reality-grounded vulnerability-inducing coding tasks for online RL rollouts, and (ii) training a reasoning-based vulnerability reward model that provides scalable and reliable security supervision. Together, these components are unified in an online RL loop to align code LLMs to generate secure and functional code. Extensive experiments demonstrate that SecCoderX achieves state-of-the-art performance, improving Effective Safety Rate (ESR) by approximately 10% over unaligned models, whereas prior methods often degrade ESR by 14-54%. We release our code, dataset and model checkpoints at https://github.com/AndrewWTY/SecCoderX.
Authors: Yuanxu Sun, Yuezhou Ma, Haixu Wu, Guanyang Zeng, Muye Chen, Jianmin Wang, Mingsheng Long
Abstract: Boundary representation (B-rep) is the industry standard for computer-aided design (CAD). While deep learning shows promise in processing B-rep models, existing methods suffer from a representation gap: continuous approaches offer analytical precision but are visually abstract, whereas discrete methods provide intuitive clarity at the expense of geometric precision. To bridge this gap, we introduce Brep2Shape, a novel self-supervised pre-training method designed to align abstract boundary representations with intuitive shape representations. Our method employs a geometry-aware task where the model learns to predict dense spatial points from parametric B\'ezier control points, enabling the network to better understand physical manifolds derived from abstract coefficients. To enhance this alignment, we propose a Dual Transformer backbone with parallel streams that independently encode surface and curve tokens to capture their distinct geometric properties. Moreover, the topology attention is integrated to model the interdependencies between surfaces and curves, thereby maintaining topological consistency. Experimental results demonstrate that Brep2Shape offers significant scalability, achieving state-of-the-art accuracy and faster convergence across various downstream tasks.
Authors: Yibin Feng, Tianqi Song, Yugin Tan, Zicheng Zhu, Yi-Chieh Lee
Abstract: Social norm interventions are used promote prosocial behaviors by highlighting prevalent actions, but their effectiveness is often limited in heterogeneous populations where shared understandings of desirable behaviors are lacking. This study explores whether multi-agent systems can establish "virtual social norms" to encourage donation behavior. We conducted an online experiment where participants interacted with a group of agents to discuss donation behaviors. Changes in perceived social norms, conformity, donation behavior, and user experience were measured pre- and postdiscussion. Results show that multi-agent interactions effectively increased perceived social norms and donation willingness. Notably, in-group agents led to stronger perceived social norms, higher conformity, and greater donation increases compared to out-group agents. Our findings demonstrate the potential of multi-agent systems for creating social norm interventions and offer insights into leveraging social identity dynamics to promote prosocial behavior in virtual environments.
Authors: Songhua Yang, Xuetao Li, Xuanye Fei, Mengde Li, Miao Li
Abstract: Effective human-robot interaction requires emotionally rich multimodal expressions, yet most humanoid robots lack coordinated speech, facial expressions, and gestures. Meanwhile, real-world deployment demands on-device solutions that can operate autonomously without continuous cloud connectivity. To bridging \underline{\textit{S}}peech, \underline{\textit{E}}motion, and \underline{\textit{M}}otion, we present \textit{SeM$^2$}, a Vision Language Model-based framework that orchestrates emotionally coherent multimodal interactions through three key components: a multimodal perception module capturing user contextual cues, a Chain-of-Thought reasoning for response planning, and a novel Semantic-Sequence Aligning Mechanism (SSAM) that ensures precise temporal coordination between verbal content and physical expressions. We implement both cloud-based and \underline{\textit{e}}dge-deployed versions (\textit{SeM$^2_e$}), with the latter knowledge distilled to operate efficiently on edge hardware while maintaining 95\% of the relative performance. Comprehensive evaluations demonstrate that our approach significantly outperforms unimodal baselines in naturalness, emotional clarity, and modal coherence, advancing socially expressive humanoid robotics for diverse real-world environments.
Authors: Weiji Xie, Jiakun Zheng, Jinrui Han, Jiyuan Shi, Weinan Zhang, Chenjia Bai, Xuelong Li
Abstract: Recent advances in humanoid whole-body motion tracking have enabled the execution of diverse and highly coordinated motions on real hardware. However, existing controllers are commonly driven either by predefined motion trajectories, which offer limited flexibility when user intent changes, or by continuous human teleoperation, which requires constant human involvement and limits autonomy. This work addresses the problem of how to drive a universal humanoid controller in a real-time and interactive manner. We present TextOp, a real-time text-driven humanoid motion generation and control framework that supports streaming language commands and on-the-fly instruction modification during execution. TextOp adopts a two-level architecture in which a high-level autoregressive motion diffusion model continuously generates short-horizon kinematic trajectories conditioned on the current text input, while a low-level motion tracking policy executes these trajectories on a physical humanoid robot. By bridging interactive motion generation with robust whole-body control, TextOp unlocks free-form intent expression and enables smooth transitions across multiple challenging behaviors such as dancing and jumping, within a single continuous motion execution. Extensive real-robot experiments and offline evaluations demonstrate instant responsiveness, smooth whole-body motion, and precise control. The project page and the open-source code are available at https://text-op.github.io/
Authors: Jinzong Dong, Wei Huang, Jianshu Zhang, Zhuo Chen, Xinzhe Yuan, Qinying Gu, Zhaohui Jiang, Nanyang Ye
Abstract: Offline reinforcement learning (RL) optimizes policies from a previously collected static dataset and is an important branch of RL. A popular and promising approach is to regularize actor-critic methods with behavior cloning (BC), which yields realistic policies and mitigates bias from out-of-distribution actions, but can impose an often-overlooked performance ceiling: when dataset actions are suboptimal, indiscriminate imitation structurally prevents the actor from fully exploiting high-value regions suggested by the critic, especially in later training when imitation is already dominant. We formally analyzed this limitation by investigating convergence properties of BC-regularized actor-critic optimization and verified it on a controlled continuous bandit task. To break this ceiling, we propose proximal action replacement (PAR), a plug-and-play training sample replacer that progressively replaces low-value actions with high-value actions generated by a stable actor, broadening the action exploration space while reducing the impact of low-value data. PAR is compatible with multiple BC regularization paradigms. Extensive experiments across offline RL benchmarks show that PAR consistently improves performance and approaches state-of-the-art when combined with the basic TD3+BC.
Authors: Qinglin Zhu, Tianyu Chen, Shuai Lu, Lei Ji, Runcong Zhao, Murong Ma, Xiangxiang Dai, Yulan He, Lin Gui, Peng cheng, Yeyun Gong
Abstract: Repository-level code editing requires models to understand complex dependencies and execute precise multi-file modifications across a large codebase. While recent gains on SWE-bench rely heavily on complex agent scaffolding, it remains unclear how much of this capability can be internalised via high-quality training signals. To address this, we propose Clean Pull Request (Clean-PR), a mid-training paradigm that leverages real-world GitHub pull requests as a training signal for repository-level editing. We introduce a scalable pipeline that converts noisy pull request diffs into Search/Replace edit blocks through reconstruction and validation, resulting in the largest publicly available corpus of 2 million pull requests spanning 12 programming languages. Using this training signal, we perform a mid-training stage followed by an agentless-aligned supervised fine-tuning process with error-driven data augmentation. On SWE-bench, our model significantly outperforms the instruction-tuned baseline, achieving absolute improvements of 13.6% on SWE-bench Lite and 12.3% on SWE-bench Verified. These results demonstrate that repository-level code understanding and editing capabilities can be effectively internalised into model weights under a simplified, agentless protocol, without relying on heavy inference-time scaffolding.
Authors: Francesco Cagnetta, Allan Ravent\'os, Surya Ganguli, Matthieu Wyart
Abstract: Despite the fact that experimental neural scaling laws have substantially guided empirical progress in large-scale machine learning, no existing theory can quantitatively predict the exponents of these important laws for any modern LLM trained on any natural language dataset. We provide the first such theory in the case of data-limited scaling laws. We isolate two key statistical properties of language that alone can predict neural scaling exponents: (i) the decay of pairwise token correlations with time separation between token pairs, and (ii) the decay of the next-token conditional entropy with the length of the conditioning context. We further derive a simple formula in terms of these statistics that predicts data-limited neural scaling exponents from first principles without any free parameters or synthetic data models. Our theory exhibits a remarkable match with experimentally measured neural scaling laws obtained from training GPT-2 and LLaMA style models from scratch on two qualitatively different benchmarks, TinyStories and WikiText.
Authors: Peizhen Li, Longbing Cao, Xiao-Ming Wu, Yang Zhang
Abstract: Humanoid facial expression shadowing enables robots to realistically imitate human facial expressions in real time, which is critical for lifelike, facially expressive humanoid robots and affective human-robot interaction. Existing progress in humanoid facial expression imitation remains limited, often failing to achieve either real-time performance or realistic expressiveness due to offline video-based inference designs and insufficient ability to capture and transfer subtle expression details. To address these limitations, we present VividFace, a real-time and realistic facial expression shadowing system for humanoid robots. An optimized imitation framework X2CNet++ enhances expressiveness by fine-tuning the human-to-humanoid facial motion transfer module and introducing a feature-adaptation training strategy for better alignment across different image sources. Real-time shadowing is further enabled by a video-stream-compatible inference pipeline and a streamlined workflow based on asynchronous I/O for efficient communication across devices. VividFace produces vivid humanoid faces by mimicking human facial expressions within 0.05 seconds, while generalizing across diverse facial configurations. Extensive real-world demonstrations validate its practical utility. Videos are available at: https://lipzh5.github.io/VividFace/.
Authors: Yuhao Wang, Shengfang Zhai, Guanghao Jin, Yinpeng Dong, Linyi Yang, Jiaheng Zhang
Abstract: Large Language Model (LLM)-based agents employ external and internal memory systems to handle complex, goal-oriented tasks, yet this exposes them to severe extraction attacks, and effective defenses remain lacking. In this paper, we propose MemPot, the first theoretically verified defense framework against memory extraction attacks by injecting optimized honeypots into the memory. Through a two-stage optimization process, MemPot generates trap documents that maximize the retrieval probability for attackers while remaining inconspicuous to benign users. We model the detection process as Wald's Sequential Probability Ratio Test (SPRT) and theoretically prove that MemPot achieves a lower average number of sampling rounds compared to optimal static detectors. Empirically, MemPot significantly outperforms state-of-the-art baselines, achieving a 50% improvement in detection AUROC and an 80% increase in True Positive Rate under low False Positive Rate constraints. Furthermore, our experiments confirm that MemPot incurs zero additional online inference latency and preserves the agent's utility on standard tasks, verifying its superiority in safety, harmlessness, and efficiency.
Authors: Shanlei Mu, Yuchen Jiang, Shikang Wu, Shiyong Hong, Tianmu Sha, Junjie Zhang, Jie Zhu, Zhe Chen, Zhe Wang, Jingjian Lin
Abstract: Industrial recommender systems increasingly adopt multi-scenario learning (MSL) and multi-task learning (MTL) to handle diverse user interactions and contexts, but existing approaches suffer from two critical drawbacks: (1) underutilization of large-scale model parameters due to limited interaction with complex feature modules, and (2) difficulty in jointly modeling scenario and task information in a unified framework. To address these challenges, we propose a unified \textbf{M}ulti-\textbf{D}istribution \textbf{L}earning (MDL) framework, inspired by the "prompting" paradigm in large language models (LLMs). MDL treats scenario and task information as specialized tokens rather than auxiliary inputs or gating signals. Specifically, we introduce a unified information tokenization module that transforms features, scenarios, and tasks into a unified tokenized format. To facilitate deep interaction, we design three synergistic mechanisms: (1) feature token self-attention for rich feature interactions, (2) domain-feature attention for scenario/task-adaptive feature activation, and (3) domain-fused aggregation for joint distribution prediction. By stacking these interactions, MDL enables scenario and task information to "prompt" and activate the model's vast parameter space in a bottom-up, layer-wise manner. Extensive experiments on real-world industrial datasets demonstrate that MDL significantly outperforms state-of-the-art MSL and MTL baselines. Online A/B testing on Douyin Search platform over one month yields +0.0626\% improvement in LT30 and -0.3267\% reduction in change query rate. MDL has been fully deployed in production, serving hundreds of millions of users daily.
Authors: Mowmita Parvin Hera, Md. Shahriar Mahmud Kallol, Shohanur Rahman Nirob, Md. Badsha Bulbul, Jubayer Ahmed, M. Zhourul Islam, Hazrat Ali, Mohammmad Farhad Bulbul
Abstract: Accurate identification of cat breeds from images is a challenging task due to subtle differences in fur patterns, facial structure, and color. In this paper, we present a deep learning-based approach for classifying cat breeds using a subset of the Oxford-IIIT Pet Dataset, which contains high-resolution images of various domestic breeds. We employed the Global Context Vision Transformer (GCViT) architecture-tiny for cat breed recognition. To improve model generalization, we used extensive data augmentation, including rotation, horizontal flipping, and brightness adjustment. Experimental results show that the GCViT-Tiny model achieved a test accuracy of 92.00% and validation accuracy of 94.54%. These findings highlight the effectiveness of transformer-based architectures for fine-grained image classification tasks. Potential applications include veterinary diagnostics, animal shelter management, and mobile-based breed recognition systems. We also provide a hugging face demo at https://huggingface.co/spaces/bfarhad/cat-breed-classifier.
URLs: https://huggingface.co/spaces/bfarhad/cat-breed-classifier.
Authors: Md Sazidur Rahman, Kjersti Engan, Kathinka D{\ae}hli Kurz, Mahdieh Khanmohammadi
Abstract: Computed tomography perfusion (CTP) at admission is routinely used to estimate the ischemic core and penumbra, while follow-up diffusion-weighted MRI (DWI) provides the definitive infarct outcome. However, single time-point segmentations fail to capture the biological heterogeneity and temporal evolution of stroke. We propose a bi-temporal analysis framework that characterizes ischemic tissue using statistical descriptors, radiomic texture features, and deep feature embeddings from two architectures (mJ-Net and nnU-Net). Bi-temporal refers to admission (T1) and post-treatment follow-up (T2). All features are extracted at T1 from CTP, with follow-up DWI aligned to ensure spatial correspondence. Manually delineated masks at T1 and T2 are intersected to construct six regions of interest (ROIs) encoding both initial tissue state and final outcome. Features were aggregated per region and analyzed in feature space. Evaluation on 18 patients with successful reperfusion demonstrated meaningful clustering of region-level representations. Regions classified as penumbra or healthy at T1 that ultimately recovered exhibited feature similarity to preserved brain tissue, whereas infarct-bound regions formed distinct groupings. Both baseline GLCM and deep embeddings showed a similar trend: penumbra regions exhibit features that are significantly different depending on final state, whereas this difference is not significant for core regions. Deep feature spaces, particularly mJ-Net, showed strong separation between salvageable and non-salvageable tissue, with a penumbra separation index that differed significantly from zero (Wilcoxon signed-rank test). These findings suggest that encoder-derived feature manifolds reflect underlying tissue phenotypes and state transitions, providing insight into imaging-based quantification of stroke evolution.
Authors: Subba Reddy Oota, Vijay Rowtula, Satya Sai Srinath Namburi, Khushbu Pahwa, Anant Khandelwal, Manish Gupta, Tanmoy Chakraborty, Bapi S. Raju
Abstract: Recent work has shown that scaling large language models (LLMs) improves their alignment with human brain activity, yet it remains unclear what drives these gains and which representational properties are responsible. Although larger models often yield better task performance and brain alignment, they are increasingly difficult to analyze mechanistically. This raises a fundamental question: what is the minimal model capacity required to capture brain-relevant representations? To address this question, we systematically investigate how constraining model scale and numerical precision affects brain alignment. We compare full-precision LLMs, small language models (SLMs), and compressed variants (quantized and pruned) by predicting fMRI responses during naturalistic language comprehension. Across model families up to 14B parameters, we find that 3B SLMs achieve brain predictivity indistinguishable from larger LLMs, whereas 1B models degrade substantially, particularly in semantic language regions. Brain alignment is remarkably robust to compression: most quantization and pruning methods preserve neural predictivity, with GPTQ as a consistent exception. Linguistic probing reveals a dissociation between task performance and brain predictivity: compression degrades discourse, syntax, and morphology, yet brain predictivity remains largely unchanged. Overall, brain alignment saturates at modest model scales and is resilient to compression, challenging common assumptions about neural scaling and motivating compact models for brain-aligned language modeling.
Authors: Hussni Mohd Zakir, Eric Tatt Wei Ho
Abstract: Recent self-supervised Vision Transformers (ViTs), such as DINOv3, provide rich feature representations for dense vision tasks. This study investigates the intrinsic few-shot semantic segmentation (FSS) capabilities of frozen DINOv3 features through a training-free baseline, FSSDINO, utilizing class-specific prototypes and Gram-matrix refinement. Our results across binary, multi-class, and cross-domain (CDFSS) benchmarks demonstrate that this minimal approach, applied to the final backbone layer, is highly competitive with specialized methods involving complex decoders or test-time adaptation. Crucially, we conduct an Oracle-guided layer analysis, identifying a significant performance gap between the standard last-layer features and globally optimal intermediate representations. We reveal a "Safest vs. Optimal" dilemma: while the Oracle proves higher performance is attainable, matching the results of compute-intensive adaptation methods, current unsupervised and support-guided selection metrics consistently yield lower performance than the last-layer baseline. This characterizes a "Semantic Selection Gap" in Foundation Models, a disconnect where traditional heuristics fail to reliably identify high-fidelity features. Our work establishes the "Last-Layer" as a deceptively strong baseline and provides a rigorous diagnostic of the latent semantic potentials in DINOv3.The code is publicly available at https://github.com/hussni0997/fssdino.
Authors: Francesco Taioli, Shiping Yang, Sonia Raychaudhuri, Marco Cristani, Unnat Jain, Angel X Chang
Abstract: Language-driven object navigation requires agents to interpret natural language descriptions of target objects, which combine intrinsic and extrinsic attributes for instance recognition and commonsense navigation. Existing methods either (i) use end-to-end trained models with vision-language embeddings, which struggle to generalize beyond training data and lack action-level explainability, or (ii) rely on modular zero-shot pipelines with large language models (LLMs) and open-set object detectors, which suffer from error propagation, high computational cost, and difficulty integrating their reasoning back into the navigation policy. To this end, we propose a compact 3B-parameter Vision-Language-Action (VLA) agent that performs human-like embodied reasoning for both object recognition and action selection, removing the need for stitched multi-model pipelines. Instead of raw embedding matching, our agent employs explicit image-grounded reasoning to directly answer "Is this the target object?" and "Why should I take this action?" The reasoning process unfolds in three stages: "think", "think summary", and "action", yielding improved explainability, stronger generalization, and more efficient navigation. Code and dataset available upon acceptance.
Authors: Antoine Gonon, Alexandre Cordonnier, Nicolas Boumal
Abstract: Match-and-copy is a core retrieval primitive used at inference time by large language models to retrieve a matching token from the context then copy its successor. Yet, understanding how this behavior emerges on natural data is challenging because retrieval and memorization are entangled. To disentangle the two, we introduce Gaussian Match-and-Copy (GMC), a minimalist benchmark that isolates long-range retrieval through pure second-order correlation signals. Numerical investigations show that this task retains key qualitative aspects of how Transformers develop match-and-copy circuits in practice, and separates architectures by their retrieval capabilities. We also analyze the optimization dynamics in a simplified attention setting. Although many solutions are a priori possible under a regression objective, including ones that do not implement retrieval, we identify an implicit-bias regime in which gradient descent drives the parameters to diverge while their direction aligns with the max-margin separator, yielding hard match selection. We prove this max-margin alignment for GD trajectories that reach vanishing empirical loss under explicit technical conditions.
Authors: Runcheng Wang, Yaru Chen, Guiguo Zhang, Honghua Jiang, Yongliang Qiao
Abstract: Precise identification of individual cows is a fundamental prerequisite for comprehensive digital management in smart livestock farming. While existing animal identification methods excel in controlled, single-camera settings, they face severe challenges regarding cross-camera generalization. When models trained on source cameras are deployed to new monitoring nodes characterized by divergent illumination, backgrounds, viewpoints, and heterogeneous imaging properties, recognition performance often degrades dramatically. This limits the large-scale application of non-contact technologies in dynamic, real-world farming environments. To address this challenge, this study proposes a cross-camera cow identification framework based on disentangled representation learning. This framework leverages the Subspace Identifiability Guarantee (SIG) theory in the context of bovine visual recognition. By modeling the underlying physical data generation process, we designed a principle-driven feature disentanglement module that decomposes observed images into multiple orthogonal latent subspaces. This mechanism effectively isolates stable, identity-related biometric features that remain invariant across cameras, thereby substantially improving generalization to unseen cameras. We constructed a high-quality dataset spanning five distinct camera nodes, covering heterogeneous acquisition devices and complex variations in lighting and angles. Extensive experiments across seven cross-camera tasks demonstrate that the proposed method achieves an average accuracy of 86.0%, significantly outperforming the Source-only Baseline (51.9%) and the strongest cross-camera baseline method (79.8%). This work establishes a subspace-theoretic feature disentanglement framework for collaborative cross-camera cow identification, offering a new paradigm for precise animal monitoring in uncontrolled smart farming environments.
Authors: Prachi Jindal, Anant Khandelwal, Manish Gupta, Bapi S. Raju, Subba Reddy Oota, Tanmoy Chakraborty
Abstract: Understanding how humans and artificial intelligence systems process complex narrative videos is a fundamental challenge at the intersection of neuroscience and machine learning. This study investigates how the temporal context length of video clips (3--12 s clips) and the narrative-task prompting shape brain-model alignment during naturalistic movie watching. Using fMRI recordings from participants viewing full-length movies, we examine how brain regions sensitive to narrative context dynamically represent information over varying timescales and how these neural patterns align with model-derived features. We find that increasing clip duration substantially improves brain alignment for multimodal large language models (MLLMs), whereas unimodal video models show little to no gain. Further, shorter temporal windows align with perceptual and early language regions, while longer windows preferentially align higher-order integrative regions, mirrored by a layer-to-cortex hierarchy in MLLMs. Finally, narrative-task prompts (multi-scene summary, narrative summary, character motivation, and event boundary detection) elicit task-specific, region-dependent brain alignment patterns and context-dependent shifts in clip-level tuning in higher-order regions. Together, our results position long-form narrative movies as a principled testbed for probing biologically relevant temporal integration and interpretable representations in long-context MLLMs.
Authors: Ruiyi Fang, Shuo Wang, Ruizhi Pu, Qiuhao Zeng, Hao Zheng, Ziyan Wang, Jiale Cai, Zhimin Mei, Song Tang, Charles Ling, Boyu Wang
Abstract: Graph Domain Adaptation (GDA) transfers knowledge from labeled source graphs to unlabeled target graphs, addressing the challenge of label scarcity. However, existing GDA methods typically assume that both source and target graphs exhibit homophily, leading existing methods to perform poorly when heterophily is present. Furthermore, the lack of labels in the target graph makes it impossible to assess its homophily level beforehand. To address this challenge, we propose a novel homophily-agnostic approach that effectively transfers knowledge between graphs with varying degrees of homophily. Specifically, we adopt a divide-and-conquer strategy that first separately reconstructs highly homophilic and heterophilic variants of both the source and target graphs, and then performs knowledge alignment separately between corresponding graph variants. Extensive experiments conducted on five benchmark datasets demonstrate the superior performance of our approach, particularly highlighting its substantial advantages on heterophilic graphs.
Authors: Jessica Ka Yi Chiu, Tom Frode Hansen, Eivind Magnus Paulsen, Ole Jakob Mengshoel
Abstract: This paper presents a geology-driven machine learning method for automated rock joint trace mapping from images. The approach combines geological modelling, synthetic data generation, and supervised image segmentation to address limited real data and class imbalance. First, discrete fracture network models are used to generate synthetic jointed rock images at field-relevant scales via parametric modelling, preserving joint persistence, connectivity, and node-type distributions. Second, segmentation models are trained using mixed training and pretraining followed by fine-tuning on real images. The method is tested in box and slope domains using several real datasets. The results show that synthetic data can support supervised joint trace detection when real data are scarce. Mixed training performs well when real labels are consistent (e.g. box-domain), while fine-tuning is more robust when labels are noisy (e.g. slope-domain where labels can be biased, incomplete, and inconsistent). Fully zero-shot prediction from synthetic model remains limited, but useful generalisation is achieved by fine-tuning with a small number of real data. Qualitative analysis shows clearer and more geologically meaningful joint traces than indicated by quantitative metrics alone. The proposed method supports reliable joint mapping and provides a basis for further work on domain adaptation and evaluation.
Authors: Yuxin Chen, Yu Wang, Yi Zhang, Ziang Ye, Zhengzhou Cai, Yaorui Shi, Qi Gu, Hui Su, Xunliang Cai, Xiang Wang, An Zhang, Tat-Seng Chua
Abstract: Recent large language models (LLMs) achieve strong performance in generating promising reasoning paths for complex tasks. However, despite powerful generation ability, LLMs remain weak at verifying their own answers, revealing a persistent capability asymmetry between generation and self-verification. In this work, we conduct an in-depth investigation of this asymmetry throughout training evolution and show that, even on the same task, improving generation does not lead to corresponding improvements in self-verification. Interestingly, we find that the reverse direction of this asymmetry behaves differently: learning to self-verify can effectively improve generation performance, achieving accuracy comparable to standard generation training while yielding more efficient and effective reasoning traces. Building on this observation, we further explore integrating self-verification into generation training by formulating a multi-task reinforcement learning framework, where generation and self-verification are optimized as two independent but complementary objectives. Extensive experiments across benchmarks and models demonstrate performance gains over generation-only training in both generation and verification capabilities.
Authors: Yuanzhi Liang, Xuan'er Wu, Yirui Liu, Yijie Fang, Yizhen Fan, Ke Hao, Rui Li, Ruiying Liu, Ziqi Ni, Peng Yu, Yanbo Wang, Haibin Huang, Qizhen Weng, Chi Zhang, Xuelong Li
Abstract: Post-training is the decisive step for converting a pretrained video generator into a production-oriented model that is instruction-following, controllable, and robust over long temporal horizons. This report presents a systematical post-training framework that organizes supervised policy shaping, reward-driven reinforcement learning, and preference-based refinement into a single stability-constrained optimization stack. The framework is designed around practical video-generation constraints, including high rollout cost, temporally compounding failure modes, and feedback that is heterogeneous, uncertain, and often weakly discriminative. By treating optimization as a staged, diagnostic-driven process rather than a collection of isolated tricks, the report summarizes a cohesive recipe for improving perceptual fidelity, temporal coherence, and prompt adherence while preserving the controllability established at initialization. The resulting framework provides a clear blueprint for building scalable post-training pipelines that remain stable, extensible, and effective in real-world deployment settings.
Authors: Xi Chen, Ming Li, Junxi Li, Changsheng Li, Peisong Wang, Lizhong Ding, Ye Yuan, Guoren Wang
Abstract: Weight-only post-training quantization (PTQ) is crucial for efficient Large Language Model (LLM) deployment but suffers from accuracy degradation caused by weight and activation outliers. Existing mitigation strategies often face critical limitations: they either yield insufficient outlier suppression or incur significant deployment inefficiencies, such as inference latency, heavy preprocessing, or reliance on complex operator fusion. To resolve these limitations, we leverage a key insight: over-parameterized LLMs often converge to Flat Minima, implying a vast equivalent solution space where weights can be adjusted without compromising accuracy. Building on this, we propose Astro, an Activation-guided Structured Regularization framework designed to suppress the negative effects of outliers in a hardware-friendly and efficient manner. Leveraging the activation-guided regularization objective, Astro actively reconstructs intrinsically robust weights, aggressively suppressing weight outliers corresponding to high-magnitude activations without sacrificing model accuracy. Crucially, Astro introduces zero inference latency and is orthogonal to mainstream quantization methods like GPTQ. Extensive experiments show that Astro achieves highly competitive performance; notably, on LLaMA-2-7B, it achieves better performance than complex learning-based rotation methods with almost 1/3 of the quantization time.
Authors: Hulingxiao He, Zijun Geng, Yuxin Peng
Abstract: Any entity in the visual world can be hierarchically grouped based on shared characteristics and mapped to fine-grained sub-categories. While Multi-modal Large Language Models (MLLMs) achieve strong performance on coarse-grained visual tasks, they often struggle with Fine-Grained Visual Recognition (FGVR). Adapting general-purpose MLLMs to FGVR typically requires large amounts of annotated data, which is costly to obtain, leaving a substantial performance gap compared to contrastive CLIP models dedicated for discriminative tasks. Moreover, MLLMs tend to overfit to seen sub-categories and generalize poorly to unseen ones. To address these challenges, we propose Fine-R1, an MLLM tailored for FGVR through an R1-style training framework: (1) Chain-of-Thought Supervised Fine-tuning, where we construct a high-quality FGVR CoT dataset with rationales of "visual analysis, candidate sub-categories, comparison, and prediction", transition the model into a strong open-world classifier; and (2) Triplet Augmented Policy Optimization, where Intra-class Augmentation mixes trajectories from anchor and positive images within the same category to improve robustness to intra-class variance, while Inter-class Augmentation maximizes the response distinction conditioned on images across sub-categories to enhance discriminative ability. With only 4-shot training, Fine-R1 outperforms existing general MLLMs, reasoning MLLMs, and even contrastive CLIP models in identifying both seen and unseen sub-categories, showing promise in working in knowledge-intensive domains where gathering expert annotations for all sub-categories is arduous. Code is available at https://github.com/PKU-ICST-MIPL/FineR1_ICLR2026.
Authors: Ruoyu Su, Alexander Bakhtin, Noman Ahmad, Matteo Esposito, Valentina Lenarduzzi, Davide Taibi
Abstract: Architectural Decision Records (ADRs) play a central role in maintaining software architecture quality, yet many decision violations go unnoticed because projects lack both systematic documentation and automated detection mechanisms. Recent advances in Large Language Models (LLMs) open up new possibilities for automating architectural reasoning at scale. We investigated how effectively LLMs can identify decision violations in open-source systems by examining their agreement, accuracy, and inherent limitations. Our study analyzed 980 ADRs across 109 GitHub repositories using a multi-model pipeline in which one LLM primary screens potential decision violations, and three additional LLMs independently validate the reasoning. We assessed agreement, accuracy, precision, and recall, and complemented the quantitative findings with expert evaluation. The models achieved substantial agreement and strong accuracy for explicit, code-inferable decisions. Accuracy falls short for implicit or deployment-oriented decisions that depend on deployment configuration or organizational knowledge. Therefore, LLMs can meaningfully support validation of architectural decision compliance; however, they are not yet replacing human expertise for decisions not focused on code.
Authors: Juntong Wu, Jialiang Cheng, Fuyu Lv, Ou Dan, Li Yuan
Abstract: Mixture-of-Experts (MoE) architectures employ sparse activation to deliver faster training and inference with higher accuracy than dense LLMs. However, in production serving, MoE models require batch inference to optimize hardware efficiency, which may cause excessive expert activation and thus slow the memory-bound decoding stage. To address the fundamental tension between batch decoding and expert sparsity, we present SERE, a Similarity-based Expert Re-routing method for Efficient batch decoding in MoE models. SERE dynamically reduces the number of active experts in an input-aware manner by re-routing tokens from secondary experts to their most similar primary counterparts. It also leverages similarity patterns to identify and preserve critical experts, thereby preventing capability loss. Notably, SERE avoids static expert pruning or merging, instead enabling dynamic expert skipping based on batch-level expert redundancy. Additionally, we provide an efficient custom CUDA kernel for SERE, enabling plug-and-play use in vLLM with only a single-line code change. Extensive experiments on various complex reasoning benchmarks demonstrate that SERE achieves up to 2.0x speedup with minimal quality loss, providing a practical solution for cost-efficient and latency-sensitive large-scale MoE deployment. Code implementation of SERE can be found in https://github.com/JL-Cheng/SERE.
Authors: Binxiao Xu, Junyu Feng, Xiaopeng Lin, Haodong Li, Zhiyuan Feng, Bohan Zeng, Shaolin Lu, Ming Lu, Qi She, Wentao Zhang
Abstract: Multimodal understanding of advertising videos is essential for interpreting the intricate relationship between visual storytelling and abstract persuasion strategies. However, despite excelling at general search, existing agents often struggle to bridge the cognitive gap between pixel-level perception and high-level marketing logic. To address this challenge, we introduce AD-MIR, a framework designed to decode advertising intent via a two-stage architecture. First, in the Structure-Aware Memory Construction phase, the system converts raw video into a structured database by integrating semantic retrieval with exact keyword matching. This approach prioritizes fine-grained brand details (e.g., logos, on-screen text) while dynamically filtering out irrelevant background noise to isolate key protagonists. Second, the Structured Reasoning Agent mimics a marketing expert through an iterative inquiry loop, decomposing the narrative to deduce implicit persuasion tactics. Crucially, it employs an evidence-based self-correction mechanism that rigorously validates these insights against specific video frames, automatically backtracking when visual support is lacking. Evaluation on the AdsQA benchmark demonstrates that AD-MIR achieves state-of-the-art performance, surpassing the strongest general-purpose agent, DVD, by 1.8% in strict and 9.5% in relaxed accuracy. These results underscore that effective advertising understanding demands explicitly grounding abstract marketing strategies in pixel-level evidence. The code is available at https://github.com/Little-Fridge/AD-MIR.
Authors: Leonardo Gonzalez
Abstract: Infographics are widely used to communicate information with a combination of text, icons, and data visualizations, but once exported as images their content is locked into pixels, making updates, localization, and reuse expensive. We describe \textsc{Images2Slides}, an API-based pipeline that converts a static infographic (PNG/JPG) into a native, editable Google Slides slide by extracting a region-level specification with a vision-language model (VLM), mapping pixel geometry into slide coordinates, and recreating elements using the Google Slides batch update API. The system is model-agnostic and supports multiple VLM backends via a common JSON region schema and deterministic postprocessing. On a controlled benchmark of 29 programmatically generated infographic slides with known ground-truth regions, \textsc{Images2Slides} achieves an overall element recovery rate of $0.989\pm0.057$ (text: $0.985\pm0.083$, images: $1.000\pm0.000$), with mean text transcription error $\mathrm{CER}=0.033\pm0.149$ and mean layout fidelity $\mathrm{IoU}=0.364\pm0.161$ for text regions and $0.644\pm0.131$ for image regions. We also highlight practical engineering challenges in reconstruction, including text size calibration and non-uniform backgrounds, and describe failure modes that guide future work.
Authors: Sai Puppala, Ismail Hossain, Md Jahangir Alam, Yoonpyo Lee, Jay Yoo, Tanzim Ahad, Syed Bahauddin Alam, Sajedul Talukder
Abstract: Large language models are increasingly deployed as *deep agents* that plan, maintain persistent state, and invoke external tools, shifting safety failures from unsafe text to unsafe *trajectories*. We introduce **AgentFence**, an architecture-centric security evaluation that defines 14 trust-boundary attack classes spanning planning, memory, retrieval, tool use, and delegation, and detects failures via *trace-auditable conversation breaks* (unauthorized or unsafe tool use, wrong-principal actions, state/objective integrity violations, and attack-linked deviations). Holding the base model fixed, we evaluate eight agent archetypes under persistent multi-turn interaction and observe substantial architectural variation in mean security break rate (MSBR), ranging from $0.29 \pm 0.04$ (LangGraph) to $0.51 \pm 0.07$ (AutoGPT). The highest-risk classes are operational: Denial-of-Wallet ($0.62 \pm 0.08$), Authorization Confusion ($0.54 \pm 0.10$), Retrieval Poisoning ($0.47 \pm 0.09$), and Planning Manipulation ($0.44 \pm 0.11$), while prompt-centric classes remain below $0.20$ under standard settings. Breaks are dominated by boundary violations (SIV 31%, WPA 27%, UTI+UTA 24%, ATD 18%), and authorization confusion correlates with objective and tool hijacking ($\rho \approx 0.63$ and $\rho \approx 0.58$). AgentFence reframes agent security around what matters operationally: whether an agent stays within its goal and authority envelope over time.
Authors: Matthew Siper, Muhammad Umair Nasir, Ahmed Khalifa, Lisa Soros, Jay Azhang, Julian Togelius
Abstract: Genetic Programming yields interpretable programs, but small syntactic mutations can induce large, unpredictable behavioral shifts, degrading locality and sample efficiency. We frame this as an operator-design problem: learn a continuous program space where latent distance has behavioral meaning, then design mutation operators that exploit this structure without changing the evolutionary optimizer. We make locality measurable by tracking action-level divergence under controlled latent perturbations, identifying an empirical trust region for behavior-local continuous variation. Using a compact trading-strategy DSL with four semantic components (long/short entry and exit), we learn a matching block-factorized embedding and compare isotropic Gaussian mutation over the full latent space to geometry-compiled mutation that restricts updates to semantically paired entry--exit subspaces and proposes directions using a learned flow-based model trained on logged mutation outcomes. Under identical $(\mu+\lambda)$ evolution strategies and fixed evaluation budgets across five assets, the learned mutation operator discovers strong strategies using an order of magnitude fewer evaluations and achieves the highest median out-of-sample Sharpe ratio. Although isotropic mutation occasionally attains higher peak performance, geometry-compiled mutation yields faster, more reliable progress, demonstrating that semantically aligned mutation can substantially improve search efficiency without modifying the underlying evolutionary algorithm.
Authors: Cen Zhang, Younggi Park, Fabian Fleischer, Yu-Fu Fu, Jiho Kim, Dongkwan Kim, Youngjoon Kim, Qingxiao Xu, Andrew Chin, Ze Sheng, Hanqing Zhao, Brian J. Lee, Joshua Wang, Michael Pelican, David J. Musliner, Jeff Huang, Jon Silliman, Mikel Mcdaniel, Jefferson Casavant, Isaac Goldthwaite, Nicholas Vidovich, Matthew Lehman, Taesoo Kim
Abstract: DARPA's AI Cyber Challenge (AIxCC, 2023--2025) is the largest competition to date for building fully autonomous cyber reasoning systems (CRSs) that leverage recent advances in AI -- particularly large language models (LLMs) -- to discover and remediate vulnerabilities in real-world open-source software. This paper presents the first systematic analysis of AIxCC. Drawing on design documents, source code, execution traces, and discussions with organizers and competing teams, we examine the competition's structure and key design decisions, characterize the architectural approaches of finalist CRSs, and analyze competition results beyond the final scoreboard. Our analysis reveals the factors that truly drove CRS performance, identifies genuine technical advances achieved by teams, and exposes limitations that remain open for future research. We conclude with lessons for organizing future competitions and broader insights toward deploying autonomous CRSs in practice.
Authors: Ross Greer, Laura Fleig, Maitrayee Keskar, Erika Maquiling, Giovanni Tapia Lopez, Angel Martinez-Sanchez, Parthib Roy, Jake Rattigan, Mira Sur, Alejandra Vidrio, Thomas Marcotte, Mohan Trivedi
Abstract: The looking-in-looking-out (LILO) framework has enabled intelligent vehicle applications that understand both the outside scene and the driver state to improve safety outcomes, with examples in smart airbag deployment, takeover time prediction in autonomous control transitions, and driver attention monitoring. In this research, we propose an augmentation to this framework, making a case for the audio modality as an additional source of information to understand the driver, and in the evolving autonomy landscape, also the passengers and those outside the vehicle. We expand LILO by incorporating audio signals, forming the looking-and-listening inside-and-outside (L-LIO) framework to enhance driver state assessment and environment understanding through multimodal sensor fusion. We evaluate three example cases where audio enhances vehicle safety: supervised learning on driver speech audio to classify potential impairment states (e.g., intoxication), collection and analysis of passenger natural language instructions (e.g., "turn after that red building") to motivate how spoken language can interface with planning systems through audio-aligned instruction data, and limitations of vision-only systems where audio may disambiguate the guidance and gestures of external agents. Datasets include custom-collected in-vehicle and external audio samples in real-world environments. Pilot findings show that audio yields safety-relevant insights, particularly in nuanced or context-rich scenarios where sound is critical to safe decision-making or visual signals alone are insufficient. Challenges include ambient noise interference, privacy considerations, and robustness across human subjects, motivating further work on reliability in dynamic real-world contexts. L-LIO augments driver and scene understanding through multimodal fusion of audio and visual sensing, offering new paths for safety intervention.
Authors: Jarrod Barnes
Abstract: Test-time training (TTT) adapts language models through gradient-based updates at inference. But is adaptation the right strategy? We study compute-optimal test-time strategies for verifiable execution-grounded (VEG) tasks, domains like GPU kernel optimization where a deterministic evaluator provides dense, continuous reward signals. Using KernelBench as our testbed and a 120B-parameter model (GPT-OSS-120B with LoRA adaptation), we find that search outperforms minimal adaptation (1-5 gradient steps): Best-of-N sampling achieves 90% task success (18/20 tasks) at K=64 across the full KernelBench L1 eval set while TTT's best checkpoint reaches only 30.6% (3-seed mean), with TTT's "equivalent K" falling below 1, worse than single-sample inference. The failure mode is over-sharpening: gradient updates collapse diversity toward mediocre solutions rather than discovering optimal ones. Our main contribution is surprisal-guided selection: selecting the highest-surprisal (lowest-confidence) correct sample yields 80% success vs. 50% for most-confident selection, a 30% improvement. Extending to surprisal-guided-top3 matches oracle performance at 100%. This zero-cost strategy, validated through length-controlled analysis, recovers oracle performance. For dense-reward VEG tasks, compute should be allocated to sample diversity and intelligent selection rather than gradient adaptation. The surprisal-guided selection principle may generalize to other execution-grounded domains where optimal solutions occupy the distribution tail.
Authors: Babak Rahmani
Abstract: Code World Models (CWMs) are language models trained to simulate program execution by predicting explicit runtime state after every executed command. This execution-based world modeling enables internal verification within the model, offering an alternative to natural language chain-of-thought reasoning. However, the sources of errors and the nature of CWMs' limitations remain poorly understood. We study CWMs from two complementary perspectives: local semantic execution and long-horizon state tracking. On real-code benchmarks, we identify two dominant failure regimes. First, dense runtime state reveals produce token-intensive execution traces, leading to token-budget exhaustion on programs with long execution histories. Second, failures disproportionately concentrate in string-valued state, which we attribute to limitations of subword tokenization rather than program structure. To study long-horizon behavior, we use a controlled permutation-tracking benchmark that isolates state propagation under action execution. We show that long-horizon degradation is driven primarily by incorrect action generation: when actions are replaced with ground-truth commands, a Transformer-based CWM propagates state accurately over long horizons, despite known limitations of Transformers in long-horizon state tracking. These findings suggest directions for more efficient supervision and state representations in CWMs that are better aligned with program execution and data types.
Authors: Jusheng Zhang, Yijia Fan, Kaitong Cai, Jing Yang, Yongsen Zheng, Kwok-Yan Lam, Liang Lin, Keze Wang
Abstract: Gating mechanisms are ubiquitous, yet a complementary question in feed-forward networks remains under-explored: how to introduce frequency-rich expressivity without sacrificing stability and scalability? This tension is exposed by spline-based Kolmogorov-Arnold Network (KAN) parameterizations, where grid refinement can induce parameter growth and brittle optimization in high dimensions. To propose a stability-preserving way to inject spectral capacity into existing MLP/FFN layers under fixed parameter and training budgets, we introduce Spectral Gating Networks (SGN), a drop-in spectral reparameterization. SGN augments a standard activation pathway with a compact spectral pathway and learnable gates that allow the model to start from a stable base behavior and progressively allocate capacity to spectral features during training. The spectral pathway is instantiated with trainable Random Fourier Features (learned frequencies and phases), replacing grid-based splines and removing resolution dependence. A hybrid GELU-Fourier formulation further improves optimization robustness while enhancing high-frequency fidelity. Across vision, NLP, audio, and PDE benchmarks, SGN consistently improves accuracy-efficiency trade-offs under comparable computational budgets, achieving 93.15% accuracy on CIFAR-10 and up to 11.7x faster inference than spline-based KAN variants. Code and trained models will be released.
Authors: Ross Greer, Maitrayee Keskar, Angel Martinez-Sanchez, Parthib Roy, Shashank Shriram, Mohan Trivedi
Abstract: Vision-language models (VLMs) have recently emerged as powerful representation learning systems that align visual observations with natural language concepts, offering new opportunities for semantic reasoning in safety-critical autonomous driving. This paper investigates how vision-language representations support driving scene safety assessment and decision-making when integrated into perception, prediction, and planning pipelines. We study three complementary system-level use cases. First, we introduce a lightweight, category-agnostic hazard screening approach leveraging CLIP-based image-text similarity to produce a low-latency semantic hazard signal. This enables robust detection of diverse and out-of-distribution road hazards without explicit object detection or visual question answering. Second, we examine the integration of scene-level vision-language embeddings into a transformer-based trajectory planning framework using the Waymo Open Dataset. Our results show that naively conditioning planners on global embeddings does not improve trajectory accuracy, highlighting the importance of representation-task alignment and motivating the development of task-informed extraction methods for safety-critical planning. Third, we investigate natural language as an explicit behavioral constraint on motion planning using the doScenes dataset. In this setting, passenger-style instructions grounded in visual scene elements suppress rare but severe planning failures and improve safety-aligned behavior in ambiguous scenarios. Taken together, these findings demonstrate that vision-language representations hold significant promise for autonomous driving safety when used to express semantic risk, intent, and behavioral constraints. Realizing this potential is fundamentally an engineering problem requiring careful system design and structured grounding rather than direct feature injection.
Authors: Qishi Zhan, Cheng-Han Yu, Yuchi Chen, Zhikang Dong, Rajarshi Guhaniyogi
Abstract: Understanding how environmental drivers relate to vegetation condition motivates spatially varying regression models, but estimating a separate coefficient surface for every predictor can yield noisy patterns and poor interpretability when many predictors are irrelevant. Motivated by MODIS vegetation index studies, we examine predictors from spectral bands, productivity and energy fluxes, observation geometry, and land surface characteristics. Because these relationships vary with canopy structure, climate, land use, and measurement conditions, methods should both model spatially varying effects and identify where predictors matter. We propose a spatially varying coefficient model where each coefficient surface uses a tensor product B-spline basis and a Bayesian group lasso prior on the basis coefficients. This prior induces predictor level shrinkage, pushing negligible effects toward zero while preserving spatial structure. Posterior inference uses Markov chain Monte Carlo and provides uncertainty quantification for each effect surface. We summarize retained effects with spatial significance maps that mark locations where the 95 percent posterior credible interval excludes zero, and we define a spatial coverage probability as the proportion of locations where the credible interval excludes zero. Simulations recover sparsity and achieve prediction. A MODIS application yields a parsimonious subset of predictors whose effect maps clarify dominant controls across landscapes.
Authors: Jusheng Zhang, Kaitong Cai, Jian Wang, Yongsen Zheng, Kwok-Yan Lam, Keze Wang
Abstract: Video understanding requires not only recognizing visual content but also performing temporally grounded, multi-step reasoning over long and noisy observations. We propose Process-of-Thought (PoT) Reasoning for Videos, a framework that makes the reasoning process explicit by structuring video inference into a sequence of lightweight, verifiable steps. PoT interleaves (i) temporal evidence selection, (ii) step-wise state updates, and (iii) constrained answer synthesis, enabling the model to progressively refine hypotheses while maintaining traceability to video evidence. The framework is designed to be model-agnostic and can be plugged into existing vision-language backbones, supporting both closed-book reasoning and evidence-augmented reasoning with external tools. We further introduce a unified representation for PoT traces that aligns intermediate decisions with temporal segments, which improves robustness to distractors and reduces hallucinated explanations. Extensive experiments on standard video reasoning tasks demonstrate that PoT consistently improves factual correctness and temporal grounding, while providing interpretable reasoning traces for diagnosis and downstream use.
Authors: Francesco Innocenti, El Mehdi Achour, Rafal Bogacz
Abstract: Predictive coding (PC) is a biologically plausible alternative to standard backpropagation (BP) that minimises an energy function with respect to network activities before updating weights. Recent work has improved the training stability of deep PC networks (PCNs) by leveraging some BP-inspired reparameterisations. However, the full scalability and theoretical basis of these approaches remains unclear. To address this, we study the infinite width and depth limits of PCNs. For linear residual networks, we show that the set of width- and depth-stable feature-learning parameterisations for PC is exactly the same as for BP. Moreover, under any of these parameterisations, the PC energy with equilibrated activities converges to the BP loss in a regime where the model width is much larger than the depth, resulting in PC computing the same gradients as BP. Experiments show that these results hold in practice for deep nonlinear networks, as long as an activity equilibrium seem to be reached. Overall, this work unifies various previous theoretical and empirical results and has potentially important implications for the scaling of PCNs.
Authors: Sagnik Mukherjee, Lifan Yuan, Pavan Jayasinha, Dilek Hakkani-T\"ur, Hao Peng
Abstract: Reinforcement learning (RL), particularly RL from verifiable reward (RLVR), has become a crucial phase of training large language models (LLMs) and a key focus of current scaling efforts. However, optimization practices in RL largely follow those of next-token prediction stages (e.g., pretraining and supervised fine-tuning), despite fundamental differences between RL and these stages highlighted by recent work. One such practice is the use of the AdamW optimizer, which is widely adopted for training large-scale transformers despite its high memory overhead. Our analysis shows that both momentum and adaptive learning rates in AdamW are less influential in RL than in SFT, leading us to hypothesize that RL benefits less from Adam-style per-parameter adaptive learning rates and momentum. Confirming this hypothesis, our experiments demonstrate that the substantially more memory-efficient SGD, which is known to perform poorly in supervised learning of large-scale transformers, matches or even outperforms AdamW in RL for LLMs. Remarkably, full fine-tuning with SGD updates fewer than 0.02% of model parameters without any sparsity-promoting regularization, more than 1000 times fewer than AdamW. Our analysis offers potential reasons for this update sparsity. These findings provide new insights into the optimization dynamics of RL in LLMs and show that RL can be substantially more parameter-efficient than previously recognized.
Authors: Siddarth Chandrasekar, Marlos C. Machado
Abstract: Across scientific disciplines, Laplacian eigenvectors serve as a fundamental basis for simplifying complex systems, from signal processing to quantum mechanics. In reinforcement learning (RL), these eigenvectors provide a natural basis for approximating reward functions; however, their use is typically limited to their linear span, which restricts expressivity in complex environments. We introduce the Laplacian Keyboard (LK), a hierarchical framework that goes beyond the linear span. LK constructs a task-agnostic library of options from these eigenvectors, forming a behavior basis guaranteed to contain the optimal policy for any reward within the linear span. A meta-policy learns to stitch these options dynamically, enabling efficient learning of policies outside the original linear constraints. We establish theoretical bounds on zero-shot approximation error and demonstrate empirically that LK surpasses zero-shot solutions while achieving improved sample efficiency compared to standard RL methods.
Authors: Sunil Madhow, Yuchen Liang, Ness Shroff, Yingbin Liang, Yu-Xiang Wang
Abstract: We study inference-time reward-guided alignment for generative models. Existing methods often rely on either architecture-specific adaptations or computationally costly inference procedures. We introduce Learnable Chernoff Baselines (LCBs) as a method for efficiently and approximately sampling from the exponentially tilted kernels that arise from KL-regularized reward alignment. Using only black-box sampling access to the pretrained model, LCBs implement a form of rejection sampling with adaptively selected acceptance probabilities, which allows fine-grained control over inference-compute scaling. We establish total-variation guarantees to the ideal aligned model, and demonstrate in both continuous and discrete diffusion settings that LCB sampling closely matches ideal rejection sampling while using substantially fewer queries to the pretrained model.
Authors: Hiren Madhu, Ngoc Bui, Ali Maatouk, Leandros Tassiulas, Smita Krishnaswamy, Menglin Yang, Sukanta Ganguly, Kiran Srinivasan, Rex Ying
Abstract: Embedding geometry plays a fundamental role in retrieval quality, yet dense retrievers for retrieval-augmented generation (RAG) remain largely confined to Euclidean space. However, natural language exhibits hierarchical structure from broad topics to specific entities that Euclidean embeddings fail to preserve, causing semantically distant documents to appear spuriously similar and increasing hallucination risk. To address these limitations, we introduce hyperbolic dense retrieval, developing two model variants in the Lorentz model of hyperbolic space: HyTE-FH, a fully hyperbolic transformer, and HyTE-H, a hybrid architecture projecting pre-trained Euclidean embeddings into hyperbolic space. To prevent representational collapse during sequence aggregation, we introduce the Outward Einstein Midpoint, a geometry-aware pooling operator that provably preserves hierarchical structure. On MTEB, HyTE-FH outperforms equivalent Euclidean baselines, while on RAGBench, HyTE-H achieves up to 29% gains over Euclidean baselines in context relevance and answer relevance using substantially smaller models than current state-of-the-art retrievers. Our analysis also reveals that hyperbolic representations encode document specificity through norm-based separation, with over 20% radial increase from general to specific concepts, a property absent in Euclidean embeddings, underscoring the critical role of geometric inductive bias in faithful RAG systems.
Authors: Tanmay Ambadkar, Sourav Panda, Shreyash Kale, Jonathan Dodge, Abhinav Verma
Abstract: Multi-objective reinforcement learning (MORL) seeks to learn policies that balance multiple, often conflicting objectives. Although a single preference-conditioned policy is the most flexible and scalable solution, existing approaches remain brittle in practice, frequently failing to recover complete Pareto fronts. We show that this failure stems from two structural issues in current methods: destructive gradient interference caused by premature scalarization and representational collapse across the preference space. We introduce $D^3PO$, a PPO-based framework that reorganizes multi-objective policy optimization to address these issues directly. $D^3PO$ preserves per-objective learning signals through a decomposed optimization pipeline and integrates preferences only after stabilization, enabling reliable credit assignment. In addition, a scaled diversity regularizer enforces sensitivity of policy behavior to preference changes, preventing collapse. Across standard MORL benchmarks, including high-dimensional and many-objective control tasks, $D^3PO$ consistently discovers broader and higher-quality Pareto fronts than prior single- and multi-policy methods, matching or exceeding state-of-the-art hypervolume and expected utility while using a single deployable policy.
Authors: Qiuming Luo, Yuebing Li, Feng Li, Chang Kong
Abstract: Distilling knowledge from large Vision-Language Models (VLMs) into lightweight networks is crucial yet challenging in Fine-Grained Visual Classification (FGVC), due to the reliance on fixed prompts and global alignment. To address this, we propose PAND (Prompt-Aware Neighborhood Distillation), a two-stage framework that decouples semantic calibration from structural transfer. First, we incorporate Prompt-Aware Semantic Calibration to generate adaptive semantic anchors. Second, we introduce a neighborhood-aware structural distillation strategy to constrain the student's local decision structure. PAND consistently outperforms state-of-the-art methods on four FGVC benchmarks. Notably, our ResNet-18 student achieves 76.09% accuracy on CUB-200, surpassing the strong baseline VL2Lite by 3.4%. Code is available at https://github.com/LLLVTA/PAND.
Authors: Mingfu Liang, Yufei Li, Jay Xu, Kavosh Asadi, Xi Liu, Shuo Gu, Kaushik Rangadurai, Frank Shyu, Shuaiwen Wang, Song Yang, Zhijing Li, Jiang Liu, Mengying Sun, Fei Tian, Xiaohan Wei, Chonglin Sun, Jacob Tao, Shike Mei, Hamed Firooz, Wenlin Chen, Luke Simon
Abstract: Recent studies increasingly explore Large Language Models (LLMs) as a new paradigm for recommendation systems due to their scalability and world knowledge. However, existing work has three key limitations: (1) most efforts focus on retrieval and ranking, while the reranking phase, critical for refining final recommendations, is largely overlooked; (2) LLMs are typically used in zero-shot or supervised fine-tuning settings, leaving their reasoning abilities, especially those enhanced through reinforcement learning (RL) and high-quality reasoning data, underexploited; (3) items are commonly represented by non-semantic IDs, creating major scalability challenges in industrial systems with billions of identifiers. To address these gaps, we propose the Generative Reasoning Reranker (GR2), an end-to-end framework with a three-stage training pipeline tailored for reranking. First, a pretrained LLM is mid-trained on semantic IDs encoded from non-semantic IDs via a tokenizer achieving $\ge$99% uniqueness. Next, a stronger larger-scale LLM generates high-quality reasoning traces through carefully designed prompting and rejection sampling, which are used for supervised fine-tuning to impart foundational reasoning skills. Finally, we apply Decoupled Clip and Dynamic sAmpling Policy Optimization (DAPO), enabling scalable RL supervision with verifiable rewards designed specifically for reranking. Experiments on two real-world datasets demonstrate GR2's effectiveness: it surpasses the state-of-the-art OneRec-Think by 2.4% in Recall@5 and 1.3% in NDCG@5. Ablations confirm that advanced reasoning traces yield substantial gains across metrics. We further find that RL reward design is crucial in reranking: LLMs tend to exploit reward hacking by preserving item order, motivating conditional verifiable rewards to mitigate this behavior and optimize reranking performance.
Authors: Zejun Zhang, Yixin Gan, Zhenchang Xing, Tian Zhang, Yi Li, Xiwei Xu, Qinghua Lu, Liming Zhu
Abstract: Coding standards are essential for maintaining consistent and high-quality code across teams and projects. Linters help developers enforce these standards by detecting code violations. However, manual linter configuration is complex and expertise-intensive, and the diversity and evolution of programming languages, coding standards, and linters lead to repetitive and maintenance-intensive configuration work. To reduce manual effort, we propose LintCFG, a domain-specific language (DSL)-driven, LLM-based compilation approach to automate linter configuration generation for coding standards, independent of programming languages, coding standards, and linters. Inspired by compiler design, we first design a DSL to express coding rules in a tool-agnostic, structured, readable, and precise manner. Then, we build linter configurations into DSL configuration instructions. For a given natural language coding standard, the compilation process parses it into DSL coding standards, matches them with the DSL configuration instructions to set configuration names, option names and values, verifies consistency between the standards and configurations, and finally generates linter-specific configurations. Experiments with Checkstyle for Java coding standard show that our approach achieves over 90% precision and recall in DSL representation, with accuracy, precision, recall, and F1-scores close to 70% (with some exceeding 70%) in fine-grained linter configuration generation. Notably, our approach outperforms baselines by over 100% in precision. A user study further shows that our approach improves developers' efficiency in configuring linters for coding standards. Finally, we demonstrate the generality of the approach by generating ESLint configurations for JavaScript coding standards, showcasing its broad applicability across other programming languages, coding standards, and linters.
Authors: Ningyu Xu, Qi Zhang, Xipeng Qiu, Xuanjing Huang
Abstract: Large language models (LLMs) exhibit emergent behaviors suggestive of human-like reasoning. While recent work has identified structured, human-like conceptual representations within these models, it remains unclear whether they functionally rely on such representations for reasoning. Here we investigate the internal processing of LLMs during in-context concept inference. Our results reveal a conceptual subspace emerging in middle to late layers, whose representational structure persists across contexts. Using causal mediation analyses, we demonstrate that this subspace is not merely an epiphenomenon but is functionally central to model predictions, establishing its causal role in inference. We further identify a layer-wise progression where attention heads in early-to-middle layers integrate contextual cues to construct and refine the subspace, which is subsequently leveraged by later layers to generate predictions. Together, these findings provide evidence that LLMs dynamically construct and use structured, latent representations in context for inference, offering insights into the computational processes underlying flexible adaptation.
Authors: Ruiqi Wang, Ruikang Liu, Runyu Chen, Haoxiang Suo, Zhiyi Peng, Zhuo Tang, Changjian Chen
Abstract: Detecting anomalies in tabular data is critical for many real-world applications, such as credit card fraud detection. With the rapid advancements in large language models (LLMs), state-of-the-art performance in tabular anomaly detection has been achieved by converting tabular data into text and fine-tuning LLMs. However, these methods randomly order columns during conversion, without considering the causal relationships between them, which is crucial for accurately detecting anomalies. In this paper, we present CausalTaD, a method that injects causal knowledge into LLMs for tabular anomaly detection. We first identify the causal relationships between columns and reorder them to align with these causal relationships. This reordering can be modeled as a linear ordering problem. Since each column contributes differently to the causal relationships, we further propose a reweighting strategy to assign different weights to different columns to enhance this effect. Experiments across more than 30 datasets demonstrate that our method consistently outperforms the current state-of-the-art methods. The code for CausalTAD is available at https://github.com/350234/CausalTAD.
Authors: Ching Lam Choi, Vighnesh Subramaniam, Phillip Isola, Antonio Torralba, Stefanie Jegelka
Abstract: Demographic skews in human preference data propagate systematic unfairness through reward models into aligned LLMs. We introduce Fairness Aware Reward Optimization (Faro), an in-processing framework that trains reward models under demographic parity, equalized odds, or counterfactual fairness constraints. We provide the first theoretical analysis of reward-level fairness in LLM alignment, establishing: (i) provable fairness certificates for Faro-trained rewards with controllable slack; a (ii) formal characterization of the accuracy-fairness trade-off induced by KL-regularized fine-tuning, proving fairness transfers from reward to policy; and the (iii) existence of a non-empty Pareto frontier. Unlike pre- and post-processing methods, Faro ensures reward models are simultaneously ordinal (ranking correctly), cardinal (calibrated), and fair. Across multiple LLMs and benchmarks, Faro significantly reduces bias and harmful generations while maintaining or improving model quality.
Authors: Wenqi Liu, Yunxiao Wang, Shijie Ma, Meng Liu, Qile Su, Tianke Zhang, Haonan Fan, Changyi Liu, Kaiyu Jiang, Jiankang Chen, Kaiyu Tang, Bin Wen, Fan Yang, Tingting Gao, Han Li, Yinwei Wei, Xuemeng Song
Abstract: In long-video understanding, conventional uniform frame sampling often fails to capture key visual evidence, leading to degraded performance and increased hallucinations. To address this, recent agentic thinking-with-videos paradigms have emerged, adopting a localize-clip-answer pipeline in which the model actively identifies relevant video segments, performs dense sampling within those clips, and then produces answers. However, existing methods remain inefficient, suffer from weak localization, and adhere to rigid workflows. To solve these issues, we propose VideoTemp-o3, a unified agentic thinking-with-videos framework that jointly models video grounding and question answering. VideoTemp-o3 exhibits strong localization capability, supports on-demand clipping, and can refine inaccurate localizations. Specifically, in the supervised fine-tuning stage, we design a unified masking mechanism that encourages exploration while preventing noise. For reinforcement learning, we introduce dedicated rewards to mitigate reward hacking. Besides, from the data perspective, we develop an effective pipeline to construct high-quality long video grounded QA data, along with a corresponding benchmark for systematic evaluation across various video durations. Experimental results demonstrate that our method achieves remarkable performance on both long video understanding and grounding.
Authors: Jiale Qian, Hao Meng, Tian Zheng, Pengcheng Zhu, Haopeng Lin, Yuhang Dai, Hanke Xie, Wenxiao Cao, Ruixuan Shang, Jun Wu, Hongmei Liu, Hanlin Wen, Jian Zhao, Zhonglin Jiang, Yong Chen, Shunshun Yin, Ming Tao, Jianguo Wei, Lei Xie, Xinsheng Wang
Abstract: While recent years have witnessed rapid progress in speech synthesis, open-source singing voice synthesis (SVS) systems still face significant barriers to industrial deployment, particularly in terms of robustness and zero-shot generalization. In this report, we introduce SoulX-Singer, a high-quality open-source SVS system designed with practical deployment considerations in mind. SoulX-Singer supports controllable singing generation conditioned on either symbolic musical scores (MIDI) or melodic representations, enabling flexible and expressive control in real-world production workflows. Trained on more than 42,000 hours of vocal data, the system supports Mandarin Chinese, English, and Cantonese and consistently achieves state-of-the-art synthesis quality across languages under diverse musical conditions. Furthermore, to enable reliable evaluation of zero-shot SVS performance in practical scenarios, we construct SoulX-Singer-Eval, a dedicated benchmark with strict training-test disentanglement, facilitating systematic assessment in zero-shot settings.
Authors: Xuan Ding, Pengyu Tong, Ranjie Duan, Yunjian Zhang, Rui Sun, Yao Zhu
Abstract: While large language models (LLMs) demonstrate impressive performance across various tasks, their deployment in real-world scenarios is still constrained by high computational demands. Layer-wise pruning, a commonly employed strategy to mitigate inference costs, can partially address this challenge. However, existing approaches generally depend on static heuristic rules and fail to account for the interdependencies among layers, thereby limiting the effectiveness of the pruning process. To this end, this paper proposes a game-theoretic framework that formulates layer pruning as a cooperative game in which each layer acts as a player and model performance serves as the utility. As computing exact Shapley values is computationally infeasible for large language models (LLMs), we propose using a lightweight surrogate network to estimate layer-wise marginal contributions. This network can predict LLM performance for arbitrary layer combinations at a low computational cost. Additionally, we employ stratified Monte Carlo mask sampling to further reduce the cost of Sharpley value estimation. This approach captures inter-layer dependencies and dynamically identifies critical layers for pruning. Extensive experiments demonstrate the consistent superiority of our method in terms of perplexity and zero-shot accuracy, achieving more efficient and effective layer-wise pruning for large language models.
Authors: Simiao Ren (Neo), Yuchen Zhou (Neo), Xingyu Shen (Neo), Kidus Zewde (Neo), Tommy Duong (Neo), George Huang (Neo), Hatsanai (Neo), Tiangratanakul (Dennis), Tsang (Dennis), Ng (Dennis), En Wei, Jiayu Xue
Abstract: As AI-generated images proliferate across digital platforms, reliable detection methods have become critical for combating misinformation and maintaining content authenticity. While numerous deepfake detection methods have been proposed, existing benchmarks predominantly evaluate fine-tuned models, leaving a critical gap in understanding out-of-the-box performance -- the most common deployment scenario for practitioners. We present the first comprehensive zero-shot evaluation of 16 state-of-the-art detection methods, comprising 23 pretrained detector variants (due to multiple released versions of certain detectors), across 12 diverse datasets, comprising 2.6~million image samples spanning 291 unique generators including modern diffusion models. Our systematic analysis reveals striking findings: (1)~no universal winner exists, with detector rankings exhibiting substantial instability (Spearman~$\rho$: 0.01 -- 0.87 across dataset pairs); (2)~a 37~percentage-point performance gap separates the best detector (75.0\% mean accuracy) from the worst (37.5\%); (3)~training data alignment critically impacts generalization, causing up to 20--60\% performance variance within architecturally identical detector families; (4)~modern commercial generators (Flux~Dev, Firefly~v4, Midjourney~v7) defeat most detectors, achieving only 18--30\% average accuracy; and (5)~we identify three systematic failure patterns affecting cross-dataset generalization. Statistical analysis confirms significant performance differences between detectors (Friedman test: $\chi^2$=121.01, $p<10^{-16}$, Kendall~$W$=0.524). Our findings challenge the ``one-size-fits-all'' detector paradigm and provide actionable deployment guidelines, demonstrating that practitioners must carefully select detectors based on their specific threat landscape rather than relying on published benchmark performance.
Authors: Charles Ye, Jasmine Cui
Abstract: What is the most brute-force way to install interpretable, controllable features into a model's activations? Controlling how LLMs internally represent concepts typically requires sophisticated methods to first identify, then intervene on the model's existing feature geometry. We bypass all of this. We finetune an LLM with a simple auxiliary loss, training 16 of its 3072 residual stream dimensions to be inert interpretability flags that simply indicate what concepts are required for generation. The model reorganizes around them anyway, learning to rely on these flags during actual generation tasks. As a result, these inert flags become genuine internal features: interpretable control switches that allow us to steer generation at inference time. Why does this work? When a feature is reliably supplied at a fixed location, gradient descent gradually eliminates redundant encodings elsewhere, and the model erodes its own alternative representations. A model's efficiency pressure is a lever - exploitable to induce interpretable, controllable representations.
Authors: Xian Wu, Kaijie Zhu, Ying Zhang, Lun Wang, Wenbo Guo
Abstract: Process rewards have been widely used in deep reinforcement learning to improve training efficiency, reduce variance, and prevent reward hacking. In LLM reasoning, existing works also explore various solutions for learning effective process reward models (PRM) with or without the help of an expert policy. However, existing methods either rely on strong assumptions about the expert policies (e.g., requiring their reward functions) or suffer intrinsic limitations (e.g., entropy collapse), resulting in weak PRMs or limited generalizability. In this paper, we introduce rePIRL, an inverse RL-inspired framework that learns effective PRMs with minimal assumptions about expert policies. Specifically, we design a dual learning process that updates the policy and the PRM interchangeably. Our learning algorithm has customized techniques to address the challenges of scaling traditional inverse RL to LLMs. We theoretically show that our proposed learning framework can unify both online and offline PRM learning methods, justifying that rePIRL can learn PRMs with minimal assumptions. Empirical evaluations on standardized math and coding reasoning datasets demonstrate the effectiveness of rePIRL over existing methods. We further show the application of our trained PRM in test-time training, test-time scaling, and providing an early signal for training hard problems. Finally, we validate our training recipe and key design choices via a detailed ablation study.
Authors: Weijiang Lv, Yaoxuan Feng, Xiaobo Xia, Jiayu Wang, Yan Jing, Wenchao Chen, Bo Chen
Abstract: Chain-of-Thought reasoning is widely used to improve the interpretability of multimodal large language models (MLLMs), yet the faithfulness of the generated reasoning traces remains unclear. Prior work has mainly focused on perceptual hallucinations, leaving reasoning level unfaithfulness underexplored. To isolate faithfulness from linguistic priors, we introduce SPD-Faith Bench, a diagnostic benchmark based on fine-grained image difference reasoning that enforces explicit visual comparison. Evaluations on state-of-the-art MLLMs reveal two systematic failure modes, perceptual blindness and perception-reasoning dissociation. We trace these failures to decaying visual attention and representation shifts in the residual stream. Guided by this analysis, we propose SAGE, a train-free visual evidence-calibrated framework that improves visual routing and aligns reasoning with perception. Our results highlight the importance of explicitly evaluating faithfulness beyond response correctness. Our benchmark and codes are available at https://github.com/Johanson-colab/SPD-Faith-Bench.
Authors: Jiaxi Liu, Yanzuo Jiang, Guibin Zhang, Zihan Zhang, Heng Chang, Zhenfei Yin, Qibing Ren, Junchi Yan
Abstract: Planning has become a central capability for contemporary agent systems in navigating complex, long-horizon tasks, yet existing approaches predominantly rely on fixed, hand-crafted planning structures that lack the flexibility to adapt to the structural diversity of open-ended problems. To address this limitation, we introduce TodoEvolve, a meta-planning paradigm that autonomously synthesizes and dynamically revises task-specific planning architectures. Specifically, we first construct PlanFactory, a modular design space that standardizes diverse planning paradigms within a unified codebase encompassing topology, initialization, adaptation, and navigation, thereby providing a common interface for heterogeneous planning patterns. Leveraging PlanFactory, we collect high-quality planning trajectories and train Todo-14B via \textit{Impedance-Guided Preference Optimization} (IGPO), a multi-objective reinforcement learning objective that encourages the generation of planning systems that are performant, stable, and token-efficient across arbitrary tasks and agent backbones. Empirical evaluations on five agentic benchmarks demonstrate that TodoEvolve consistently surpasses carefully engineered planning modules while maintaining economical API costs and runtime overhead.
Authors: Benjamin Le, Xueying Lu, Nick Stern, Wenqiong Liu, Igor Lapchuk, Xiang Li, Baofen Zheng, Kevin Rosenberg, Jiewen Huang, Zhe Zhang, Abraham Cabangbang, Satej Milind Wagle, Jianqiang Shen, Raghavan Muthuregunathan, Abhinav Gupta, Mathew Teoh, Andrew Kirk, Thomas Kwan, Jingwei Wu, Wenjing Zhang
Abstract: Evaluating relevance in large-scale search systems is fundamentally constrained by the governance gap between nuanced, resource-constrained human oversight and the high-throughput requirements of production systems. While traditional approaches rely on engagement proxies or sparse manual review, these methods often fail to capture the full scope of high-impact relevance failures. We present \textbf{SAGE} (Scalable AI Governance \& Evaluation), a framework that operationalizes high-quality human product judgment as a scalable evaluation signal. At the core of SAGE is a bidirectional calibration loop where natural-language \emph{Policy}, curated \emph{Precedent}, and an \emph{LLM Surrogate Judge} co-evolve. SAGE systematically resolves semantic ambiguities and misalignments, transforming subjective relevance judgment into an executable, multi-dimensional rubric with near human-level agreement. To bridge the gap between frontier model reasoning and industrial-scale inference, we apply teacher-student distillation to transfer high-fidelity judgments into compact student surrogates at \textbf{92$\times$} lower cost. Deployed within LinkedIn Search ecosystems, SAGE guided model iteration through simulation-driven development, distilling policy-aligned models for online serving and enabling rapid offline evaluation. In production, it powered policy oversight that measured ramped model variants and detected regressions invisible to engagement metrics. Collectively, these drove a \textbf{0.25\%} lift in LinkedIn daily active users.
Authors: Satyam Kumar Navneet, Joydeep Chandra, Yong Zhang
Abstract: Adaptive learning systems optimize content delivery based on performance metrics but ignore the dynamic attention fluctuations that characterize neurodivergent learners. We present AttentionGuard, a framework that detects engagement-attention states from privacy-preserving behavioral signals and adapts interface elements accordingly. Our approach models four attention states derived from ADHD phenomenology and implements five novel UI adaptation patterns including bi-directional scaffolding that responds to both understimulation and overstimulation. We validate our detection model on the OULAD dataset, achieving 87.3% classification accuracy, and demonstrate correlation with clinical ADHD profiles through cross-validation on the HYPERAKTIV dataset. A Wizard-of-Oz study with 11 adults showing ADHD characteristics found significantly reduced cognitive load in the adaptive condition (NASA-TLX: 47.2 vs 62.8, Cohen's d=1.21, p=0.008) and improved comprehension (78.4% vs 61.2%, p=0.009). Concordance analysis showed 84% agreement between wizard decisions and automated classifier predictions, supporting deployment feasibility. The system is presented as an interactive demo where observers can inspect detected attention states, observe real-time UI adaptations, and compare automated decisions with human-in-the-loop overrides. We contribute empirically validated UI patterns for attention-adaptive interfaces and evidence that behavioral attention detection can meaningfully support neurodivergent learning experiences.
Authors: Donghyeon Ki, Hee-Jun Ahn, Kyungyoon Kim, Byung-Jun Lee
Abstract: Soft policies in reinforcement learning define policies as Boltzmann distributions over state-action value functions, providing a principled mechanism for balancing exploration and exploitation. However, realizing such soft policies in practice remains challenging. Existing approaches either depend on parametric policies with limited expressivity or employ diffusion-based policies whose intractable likelihoods hinder reliable entropy estimation in soft policy objectives. We address this challenge by directly realizing soft-policy sampling via Langevin dynamics driven by the action gradient of the Q-function. This perspective leads to Langevin Q-Learning (LQL), which samples actions from the target Boltzmann distribution without explicitly parameterizing the policy. However, directly applying Langevin dynamics suffers from slow mixing in high-dimensional and non-convex Q-landscapes, limiting its practical effectiveness. To overcome this, we propose Noise-Conditioned Langevin Q-Learning (NC-LQL), which integrates multi-scale noise perturbations into the value function. NC-LQL learns a noise-conditioned Q-function that induces a sequence of progressively smoothed value landscapes, enabling sampling to transition from global exploration to precise mode refinement. On OpenAI Gym MuJoCo benchmarks, NC-LQL achieves competitive performance compared to state-of-the-art diffusion-based methods, providing a simple yet powerful solution for online RL.
Authors: Tianyi Wang, Huawei Fan, Yuanchao Shu, Peng Cheng, Cong Wang
Abstract: Large Language Models face an emerging and critical threat known as latency attacks. Because LLM inference is inherently expensive, even modest slowdowns can translate into substantial operating costs and severe availability risks. Recently, a growing body of research has focused on algorithmic complexity attacks by crafting inputs to trigger worst-case output lengths. However, we report a counter-intuitive finding that these algorithmic latency attacks are largely ineffective against modern LLM serving systems. We reveal that system-level optimization such as continuous batching provides a logical isolation to mitigate contagious latency impact on co-located users. To this end, in this paper, we shift the focus from the algorithm to the system layer, and introduce a new Fill and Squeeze attack strategy targeting the state transition of the scheduler. "Fill" first exhausts the global KV cache to induce Head-of-Line blocking, while "Squeeze" forces the system into repetitive preemption. By manipulating output lengths using methods from simple plain-text prompts to more complex prompt engineering, and leveraging side-channel probing of memory status, we demonstrate that the attack can be orchestrated in a black-box setting with much less cost. Extensive evaluations indicate by up to 20-280x average slowdown on Time to First Token and 1.5-4x average slowdown on Time Per Output Token compared to existing attacks with 30-40% lower attack cost.
Authors: Yu Ding, Yulin Shao
Abstract: Deep learning has enabled significant advances in feedback-based channel coding, yet existing learned schemes remain fundamentally limited: they employ fixed block lengths, suffer degraded performance at high rates, and cannot fully exploit the adaptive potential of feedback. This paper introduces Deep Variable-Length Feedback (DeepVLF) coding, a flexible coding framework that dynamically adjusts transmission length via learned feedback. We propose two complementary architectures: DeepVLF-R, where termination is receiver-driven, and DeepVLF-T, where the transmitter controls termination. Both architectures leverage bit-group partitioning and transformer-based encoder-decoder networks to enable fine-grained rate adaptation in response to feedback. Evaluations over AWGN and 5G-NR fading channels demonstrate that DeepVLF substantially outperforms state-of-the-art learned feedback codes. It achieves the same block error rate with 20%-55% fewer channel uses and lowers error floors by orders of magnitude, particularly in high-rate regimes. Encoding dynamics analysis further reveals that the models autonomously learn a two-phase strategy analogous to classical Schalkwijk-Kailath coding: an initial information-carrying phase followed by a noise-cancellation refinement phase. This emergent behavior underscores the interpretability and information-theoretic alignment of the learned codes.
Authors: Mohammad Ashhad, Robert Hoehndorf, Ricardo Henao
Abstract: Survival analysis is complicated by censored data, high-dimensional features, and non-linear interactions. Classical models are interpretable but restrictive, while deep learning models are flexible but often non-interpretable and sensitive to noise. We propose GRAFT (Gated Residual Accelerated Failure Time), a novel AFT model that decouples prognostic ranking from calibration. GRAFT's hybrid architecture combines a linear AFT model with a non-linear residual neural network, and it also integrates stochastic gates for automatic, end-to-end feature selection. The model is trained by directly optimizing a differentiable, C-index-aligned ranking loss using stochastic conditional imputation from local Kaplan-Meier estimators. In public benchmarks, GRAFT outperforms baselines in discrimination and calibration, while remaining robust and sparse in high-noise settings.
Authors: Enhao Chen, Yulin Shao
Abstract: This paper reimagines the foundational feedback mechanism in wireless communication, transforming the prevailing 1-bit binary ACK/NACK with a high-dimensional, information-rich vector to transform passive acknowledgment into an active collaboration. We present Rich-ARQ, a paradigm that introduces neural-coded feedback for collaborative physical-layer channel coding between transmitter and receiver. To realize this vision in practice, we develop a novel asynchronous feedback code that eliminates stalling from feedback delays, adapts dynamically to channel fluctuations, and features a lightweight encoder suitable for on-device deployment. We materialize this concept into the first full-stack, standard-compliant software-defined radio prototype, which decouples AI inference from strict radio timing. Comprehensive over-the-air experiments demonstrate that Rich-ARQ achieves significant SNR gains over conventional 1-bit hybrid ARQ and remarkable latency reduction over prior learning-based feedback codes, moving the promise of intelligent feedback from theory to a practical, high-performance reality for next-generation networks.
Authors: Zihui Gao, Ke Liu, Donny Y. Chen, Duochao Shi, Guosheng Lin, Hao Chen, Chunhua Shen
Abstract: Geometric foundation models show promise in 3D reconstruction, yet their progress is severely constrained by the scarcity of diverse, large-scale 3D annotations. While Internet videos offer virtually unlimited raw data, utilizing them as a scaling source for geometric learning is challenging due to the absence of ground-truth geometry and the presence of observational noise. To address this, we propose SAGE, a framework for Scalable Adaptation of GEometric foundation models from raw video streams. SAGE leverages a hierarchical mining pipeline to transform videos into training trajectories and hybrid supervision: (1) Informative training trajectory selection; (2) Sparse Geometric Anchoring via SfM point clouds for global structural guidance; and (3) Dense Differentiable Consistency via 3D Gaussian rendering for multi-view constraints. To prevent catastrophic forgetting, we introduce a regularization strategy using anchor data. Extensive experiments show that SAGE significantly enhances zero-shot generalization, reducing Chamfer Distance by 20-42% on unseen benchmarks (7Scenes, TUM-RGBD, Matterport3D) compared to state-of-the-art baselines. To our knowledge, SAGE pioneers the adaptation of geometric foundation models via Internet video, establishing a scalable paradigm for general-purpose 3D learning.
Authors: Zhi Chen, Zhensu Sun, Yuling Shi, Chao Peng, Xiaodong Gu, David Lo, Lingxiao Jiang
Abstract: Large Language Model (LLM) code agents increasingly resolve repository-level issues by iteratively editing code, invoking tools, and validating candidate patches. In these workflows, agents often write tests on the fly, a paradigm adopted by many high-ranking agents on the SWE-bench leaderboard. However, we observe that GPT-5.2, which writes almost no new tests, can even achieve performance comparable to top-ranking agents. This raises the critical question: whether such tests meaningfully improve issue resolution or merely mimic human testing practices while consuming a substantial interaction budget. To reveal the impact of agent-written tests, we present an empirical study that analyzes agent trajectories across six state-of-the-art LLMs on SWE-bench Verified. Our results show that while test writing is commonly adopted, but resolved and unresolved tasks within the same model exhibit similar test-writing frequencies Furthermore, these tests typically serve as observational feedback channels, where agents prefer value-revealing print statements significantly more than formal assertion-based checks. Based on these insights, we perform a controlled experiment by revising the prompts of four agents to either increase or reduce test writing. The results suggest that changes in the volume of agent-written tests do not significantly change final outcomes. Taken together, our study reveals that current test-writing practices may provide marginal utility in autonomous software engineering tasks.
Authors: Mark Griguletskii, Danil Belov, Pavel Osinenko
Abstract: Modern autonomous vehicles and robots utilize versatile sensors for localization and mapping. The fidelity of these maps is paramount, as an accurate environmental representation is a prerequisite for stable and precise localization. Factor graphs provide a powerful approach for sensor fusion, enabling the estimation of the maximum a posteriori solution. However, the discrete nature of graph-based representations, combined with asynchronous sensor measurements, complicates consistent state estimation. The design of an optimal factor graph topology remains an open challenge, especially in multi-sensor systems with asynchronous data. Conventional approaches rely on a rigid graph structure, which becomes inefficient with sensors of disparate rates. Although preintegration techniques can mitigate this for high-rate sensors, their applicability is limited. To address this problem, this work introduces a novel approach that incrementally constructs connected factor graphs, ensuring the incorporation of all available sensor data by choosing the optimal graph topology based on the external evaluation criteria. The proposed methodology facilitates graph compression, reducing the number of nodes (optimized variables) by ~30% on average while maintaining map quality at a level comparable to conventional approaches.
Authors: Giang Ngo, Dat Phan Trong, Dang Nguyen, Sunil Gupta, Svetha Venkatesh
Abstract: Bayesian Optimization critically depends on the choice of acquisition function, but no single strategy is universally optimal; the best choice is non-stationary and problem-dependent. Existing adaptive portfolio methods often base their decisions on past function values while ignoring richer information like remaining budget or surrogate model characteristics. To address this, we introduce LMABO, a novel framework that casts a pre-trained Large Language Model (LLM) as a zero-shot, online strategist for the BO process. At each iteration, LMABO uses a structured state representation to prompt the LLM to select the most suitable acquisition function from a diverse portfolio. In an evaluation across 50 benchmark problems, LMABO demonstrates a significant performance improvement over strong static, adaptive portfolio, and other LLM-based baselines. We show that the LLM's behavior is a comprehensive strategy that adapts to real-time progress, proving its advantage stems from its ability to process and synthesize the complete optimization state into an effective, adaptive policy.
Authors: Yuzhu Cai, Zexi Liu, Xinyu Zhu, Cheng Wang, Jiaao Chen, Hanrui Wang, Wei-Chen Wang, Di Jin, Siheng Chen
Abstract: Autonomous Machine Learning Engineering (MLE) requires agents to perform sustained, iterative optimization over long horizons. While recent LLM-based agents show promise, current prompt-based agents for MLE suffer from behavioral stagnation due to frozen parameters. Although Reinforcement Learning (RL) offers a remedy, applying it to MLE is hindered by prohibitive execution latency and inefficient data selection. Recognizing these challenges, we propose AceGRPO with two core components: (1) Evolving Data Buffer that continuously repurposes execution traces into reusable training tasks, and (2) Adaptive Sampling guided by a Learnability Potential function, which dynamically prioritizes tasks at the agent's learning frontier to maximize learning efficiency. Leveraging AceGRPO, our trained Ace-30B model achieves a 100% valid submission rate on MLE-Bench-Lite, approaches the performance of proprietary frontier models, and outperforms larger open-source baselines (e.g., DeepSeek-V3.2), demonstrating robust capability for sustained iterative optimization. Code is available at https://github.com/yuzhu-cai/AceGRPO.
Authors: Huiyang Yi, Xiaojian Shen, Yonggang Wu, Duxin Chen, He Wang, Wenwu Yu
Abstract: Causal discovery from time series is a fundamental task in machine learning. However, its widespread adoption is hindered by a reliance on untestable causal assumptions and by the lack of robustness-oriented evaluation in existing benchmarks. To address these challenges, we propose CausalCompass, a flexible and extensible benchmark suite designed to assess the robustness of time-series causal discovery (TSCD) methods under violations of modeling assumptions. To demonstrate the practical utility of CausalCompass, we conduct extensive benchmarking of representative TSCD algorithms across eight assumption-violation scenarios. Our experimental results indicate that no single method consistently attains optimal performance across all settings. Nevertheless, the methods exhibiting superior overall performance across diverse scenarios are almost invariably deep learning-based approaches. We further provide hyperparameter sensitivity analyses to deepen the understanding of these findings. We also find, somewhat surprisingly, that NTS-NOTEARS relies heavily on standardized preprocessing in practice, performing poorly in the vanilla setting but exhibiting strong performance after standardization. Finally, our work aims to provide a comprehensive and systematic evaluation of TSCD methods under assumption violations, thereby facilitating their broader adoption in real-world applications. The code and datasets are available at https://github.com/huiyang-yi/CausalCompass.
Authors: Nur Ahmad Khatim, Mansur Arief
Abstract: Securing petroleum infrastructure requires balancing autonomous system efficiency with human judgment for threat escalation, a challenge unaddressed by classical facility location models assuming homogeneous resources. This paper formulates the Human-Robot Co-Dispatch Facility Location Problem (HRCD-FLP), a capacitated facility location variant incorporating tiered infrastructure criticality, human-robot supervision ratio constraints, and minimum utilization requirements. We evaluate command center selection across three technology maturity scenarios. Results show transitioning from conservative (1:3 human-robot supervision) to future autonomous operations (1:10) yields significant cost reduction while maintaining complete critical infrastructure coverage. For small problems, exact methods dominate in both cost and computation time; for larger problems, the proposed heuristic achieves feasible solutions in under 3 minutes with approximately 14% optimality gap where comparison is possible. From systems perspective, our work demonstrate that optimized planning for human-robot teaming is key to achieve both cost-effective and mission-reliable deployments.
Authors: Ziyun Li, Huancheng Hu, Soon Hoe Lim, Xuyu Li, Fei Gao, Enmao Diao, Zezhen Ding, Michalis Vazirgiannis, Henrik Bostrom
Abstract: Flow-based generative models can be viewed through a physics lens: sampling transports a particle from noise to data by integrating a time-varying velocity field, and each sample corresponds to a trajectory with its own dynamical effort. Motivated by classical mechanics, we introduce Kinetic Path Energy (KPE), an action-like, per-sample diagnostic that measures the accumulated kinetic effort along an Ordinary Differential Equation (ODE) trajectory. Empirically, KPE exhibits two robust correspondences: (i) higher KPE predicts stronger semantic fidelity; (ii) high-KPE trajectories terminate on low-density manifold frontiers. We further provide theoretical guarantees linking trajectory energy to data density. Paradoxically, this correlation is non-monotonic. At sufficiently high energy, generation can degenerate into memorization. Leveraging the closed-form of empirical flow matching, we show that extreme energies drive trajectories toward near-copies of training examples. This yields a Goldilocks principle and motivates Kinetic Trajectory Shaping (KTS), a training-free two-phase inference strategy that boosts early motion and enforces a late-time soft landing, reducing memorization and improving generation quality across benchmark tasks.
Authors: Krzysztof Wr\'obel, Jan Maria Kowalski, Jerzy Surma, Igor Ciuciura, Maciej Szyma\'nski
Abstract: As Large Language Models (LLMs) become increasingly deployed in Polish language applications, the need for efficient and accurate content safety classifiers has become paramount. We present Bielik Guard, a family of compact Polish language safety classifiers comprising two model variants: a 0.1B parameter model based on MMLW-RoBERTa-base and a 0.5B parameter model based on PKOBP/polish-roberta-8k. Fine-tuned on a community-annotated dataset of 6,885 Polish texts, these models classify content across five safety categories: Hate/Aggression, Vulgarities, Sexual Content, Crime, and Self-Harm. Our evaluation demonstrates that both models achieve strong performance on multiple benchmarks. The 0.5B variant offers the best overall discrimination capability with F1 scores of 0.791 (micro) and 0.785 (macro) on the test set, while the 0.1B variant demonstrates exceptional efficiency. Notably, Bielik Guard 0.1B v1.1 achieves superior precision (77.65\%) and very low false positive rate (0.63\%) on real user prompts, outperforming HerBERT-PL-Guard (31.55\% precision, 4.70\% FPR) despite identical model size. The models are publicly available and designed to provide appropriate responses rather than simple content blocking, particularly for sensitive categories like self-harm.
Authors: Yumin Kim, Hyeonsu Lyu, Minjae Lee, Hyun Jong Yang
Abstract: Large language models (LLMs) offer significant potential for intelligent mobile services but are computationally intensive for resource-constrained devices. Mobile edge computing (MEC) allows such devices to offload inference tasks to edge servers (ESs), yet introduces latency due to communication and serverside queuing, especially in multi-user environments. In this work, we propose an uncertainty-aware offloading framework that dynamically decides whether to perform inference locally or offload it to the ES, based on token-level uncertainty and resource constraints. We define a margin-based token-level uncertainty metric and demonstrate its correlation with model accuracy. Leveraging this metric, we design a greedy offloading algorithm (GOA) that minimizes delay while maintaining accuracy by prioritizing offloading for highuncertainty queries. Our experiments show that GOA consistently achieves a favorable trade-off, outperforming baseline strategies in both accuracy and latency across varying user densities, and operates with practical computation time. These results establish GOA as a scalable and effective solution for LLM inference in MEC environments.
Authors: Vaibhav Shukla, Hardik Sharma, Adith N Reganti, Soham Wasmatkar, Bagesh Kumar, Vrijendra Singh
Abstract: Most safety evaluations of large language models (LLMs) remain anchored in English. Translation is often used as a shortcut to probe multilingual behavior, but it rarely captures the full picture, especially when harmful intent or structure morphs across languages. Some types of harm survive translation almost intact, while others distort or disappear. To study this effect, we introduce CompositeHarm, a translation-based benchmark designed to examine how safety alignment holds up as both syntax and semantics shift. It combines two complementary English datasets, AttaQ, which targets structured adversarial attacks, and MMSafetyBench, which covers contextual, real-world harms, and extends them into six languages: English, Hindi, Assamese, Marathi, Kannada, and Gujarati. Using three large models, we find that attack success rates rise sharply in Indic languages, especially under adversarial syntax, while contextual harms transfer more moderately. To ensure scalability and energy efficiency, our study adopts lightweight inference strategies inspired by edge-AI design principles, reducing redundant evaluation passes while preserving cross-lingual fidelity. This design makes large-scale multilingual safety testing both computationally feasible and environmentally conscious. Overall, our results show that translated benchmarks are a necessary first step, but not a sufficient one, toward building grounded, resource-aware, language-adaptive safety systems.
Authors: Pablo Hidalgo, Daniel Rodriguez
Abstract: In many machine learning contexts, tasks are often treated as interconnected components with the goal of leveraging knowledge transfer between them, which is the central aim of Multi-Task Learning (MTL). Consequently, this multi-task scenario requires addressing critical questions: which tasks are similar, and how and why do they exhibit similarity? In this work, we propose a multi-task similarity measure based on Explainable Artificial Intelligence (XAI) techniques, specifically Accumulated Local Effects (ALE) curves. ALE curves are compared using the Fr\'echet distance, weighted by the data distribution, and the resulting similarity measure incorporates the importance of each feature. The measure is applicable in both single-task learning scenarios, where each task is trained separately, and multi-task learning scenarios, where all tasks are learned simultaneously. The measure is model-agnostic, allowing the use of different machine learning models across tasks. A scaling factor is introduced to account for differences in predictive performance across tasks, and several recommendations are provided for applying the measure in complex scenarios. We validate this measure using four datasets, one synthetic dataset and three real-world datasets. The real-world datasets include a well-known Parkinson's dataset and a bike-sharing usage dataset -- both structured in tabular format -- as well as the CelebA dataset, which is used to evaluate the application of concept bottleneck encoders in a multitask learning setting. The results demonstrate that the measure aligns with intuitive expectations of task similarity across both tabular and non-tabular data, making it a valuable tool for exploring relationships between tasks and supporting informed decision-making.
Authors: Zheyuan Hu, Weitao Chen, Cengiz \"Oztireli, Chenliang Zhou, Fangcheng Zhong
Abstract: Partial Differential Equations are precise in modelling the physical, biological and graphical phenomena. However, the numerical methods suffer from the curse of dimensionality, high computation costs and domain-specific discretization. We aim to explore pros and cons of different PDE solvers, and apply them to specific scientific simulation problems, including forwarding solution, inverse problems and equations discovery. In particular, we extend the recent CNF (NeurIPS 2023) framework solver to multi-dependent-variable and non-linear settings, together with down-stream applications. The outcomes include implementation of selected methods, self-tuning techniques, evaluation on benchmark problems and a comprehensive survey of neural PDE solvers and scientific simulation applications.
Authors: Xuehai Bai, Xiaoling Gu, Akide Liu, Hangjie Yuan, YiFan Zhang, Jack Ma
Abstract: Recent advances in instruction-based image editing have shown remarkable progress. However, existing methods remain limited to relatively simple editing operations, hindering real-world applications that require complex and compositional instructions. In this work, we address these limitations from the perspectives of architectural design, data, and evaluation protocols. Specifically, we identify two key challenges in current models: insufficient instruction compliance and background inconsistency. To this end, we propose MCIE-E1, a Multimodal Large Language Model-Driven Complex Instruction Image Editing method that integrates two key modules: a spatial-aware cross-attention module and a background-consistent cross-attention module. The former enhances instruction-following capability by explicitly aligning semantic instructions with spatial regions through spatial guidance during the denoising process, while the latter preserves features in unedited regions to maintain background consistency. To enable effective training, we construct a dedicated data pipeline to mitigate the scarcity of complex instruction-based image editing datasets, combining fine-grained automatic filtering via a powerful MLLM with rigorous human validation. Finally, to comprehensively evaluate complex instruction-based image editing, we introduce CIE-Bench, a new benchmark with two new evaluation metrics. Experimental results on CIE-Bench demonstrate that MCIE-E1 consistently outperforms previous state-of-the-art methods in both quantitative and qualitative assessments, achieving a 23.96% improvement in instruction compliance.
Authors: Yigit Turkmen, Baturalp Buyukates, Melih Bastopcu
Abstract: Large language models (LLMs) are often ensembled together to improve overall reliability and robustness, but in practice models are strongly correlated. This raises a fundamental question: which models should be selected when forming an LLM ensemble? We formulate budgeted ensemble selection as maximizing the mutual information between the true label and predictions of the selected models. Furthermore, to explain why performance can saturate even with many models, we model the correlated errors of the models using Gaussian-copula and show an information-theoretic error floor for the performance of the ensemble. Motivated by these, we propose a simple greedy mutual-information selection algorithm that estimates the required information terms directly from data and iteratively builds an ensemble under a query budget. We test our approach in two question answering datasets and one binary sentiment classification dataset: MEDMCQA, MMLU, and IMDB movie reviews. Across all datasets, we observe that our method consistently outperforms strong baselines under the same query budget.
Authors: Jitai Hao, Qiang Huang, Yaowei Wang, Min Zhang, Jun Yu
Abstract: The deployment of efficient long-context LLMs in applications like autonomous agents, long-chain reasoning, and creative writing is fundamentally bottlenecked by the linear growth of KV cache memory. Existing compression and eviction methods often struggle to balance accuracy, compression ratio, and hardware efficiency. We propose DeltaKV, a residual-based KV cache compression framework motivated by two empirical findings: long-range inter-token similarity and highly shared latent components in KV representations. Instead of discarding tokens, DeltaKV encodes semantic residuals relative to retrieved historical references, preserving fidelity while substantially reducing storage. To translate compression gains into real system speedups, we further introduce Sparse-vLLM, a high-performance inference engine with decoupled memory management and kernels optimized for sparse and irregular KV layouts. Experiments show that DeltaKV reduces KV cache memory to 29\% of the original while maintaining near-lossless accuracy on LongBench, SCBench, and AIME. When integrated with Sparse-vLLM, it achieves up to 2$\times$ throughput improvement over vLLM in long-context scenarios, demonstrating a practical path toward scalable long-context LLM deployment. Code, model checkpoints, and datasets are available at https://github.com/CURRENTF/Sparse-vLLM.
Authors: Riya Mohan, Juana Valeria Hurtado, Rohit Mohan, Abhinav Valada
Abstract: Autonomous driving requires forecasting both geometry and semantics over time to effectively reason about future environment states. Existing vision-based occupancy forecasting methods focus on motion-related categories such as static and dynamic objects, while semantic information remains largely absent. Recent semantic occupancy forecasting approaches address this gap but rely on past occupancy predictions obtained from separate networks. This makes current methods sensitive to error accumulation and prevents learning spatio-temporal features directly from images. In this work, we present ForecastOcc, the first framework for vision-based semantic occupancy forecasting that jointly predicts future occupancy states and semantic categories. Our framework yields semantic occupancy forecasts for multiple horizons directly from past camera images, without relying on externally estimated maps. We evaluate ForecastOcc in two complementary settings: multi-view forecasting on the Occ3D-nuScenes dataset and monocular forecasting on SemanticKITTI, where we establish the first benchmark for this task. We introduce the first baselines by adapting two 2D forecasting modules within our framework. Importantly, we propose a novel architecture that incorporates a temporal cross-attention forecasting module, a 2D-to-3D view transformer, a 3D encoder for occupancy prediction, and a semantic occupancy head for voxel-level forecasts across multiple horizons. Extensive experiments on both datasets show that ForecastOcc consistently outperforms baselines, yielding semantically rich, future-aware predictions that capture scene dynamics and semantics critical for autonomous driving.
Authors: Sizhe Dang, Jiaqi Shao, Xiaodong Zheng, Guang Dai, Yan Song, Haishan Ye
Abstract: As foundation models continue to scale, pretraining increasingly relies on data-parallel distributed optimization, making bandwidth-limited gradient synchronization a key bottleneck. Orthogonally, projection-based low-rank optimizers were mainly designed for memory efficiency, but remain suboptimal for communication-limited training: one-sided synchronization still transmits an $O(rn)$ object for an $m\times n$ matrix gradient and refresh steps can dominate peak communicated bytes. We propose TSR, which brings two-sided low-rank communication to Adam-family updates (TSR-Adam) by synchronizing a compact core $U^\top G V\in\mathbb{R}^{r\times r}$, reducing the dominant per-step payload from $O(mn)$ to $O(r^2)$ while keeping moment states in low-dimensional cores. To further reduce the peak communication from subspace refresh, TSR-Adam adopts a randomized SVD-based refresh that avoids full-gradient synchronization. We additionally extend low-rank communication to embedding gradients with embedding-specific ranks and refresh schedules, yielding additional communication and memory savings over keeping embeddings dense. Across pretraining from 60M to 1B model scales, TSR-Adam reduces average communicated bytes per step by $13\times$, and on GLUE fine-tuning it reduces communication by $25\times$, while achieving comparable performance; we further provide a theoretical stationarity analysis for the proposed update. Code is available at https://github.com/DKmiyan/TSR-Adam.
Authors: Sadegh Sohani, Salar Ghazi, Farnaz Kamranfar, Sahar Pilehvar Moakhar, Mohammad Allahbakhsh, Haleh Amintoosi, Kaiwen Zhang
Abstract: This paper addresses the critical challenge of access control in modern supply chains, which operate across multiple independent and competing organizations. Existing access control is static and centralized, unable to adapt to insider threats or evolving contexts. Blockchain improves decentralization but lacks behavioral intelligence, while centralized machine learning for anomaly detection requires aggregating sensitive data, violating privacy. The proposed solution is ICBAC, an intelligent contract-based access control framework. It integrates permissioned blockchain (Hyperledger Fabric) with federated learning (FL). Built on Fabric, ICBAC uses a multi-channel architecture and three smart contracts for asset management, baseline access control, and dynamic revocation. To counter insider misuse, each channel deploys an AI agent that monitors activity and dynamically restricts access for anomalies. Federated learning allows these agents to collaboratively improve detection models without sharing raw data. For heterogeneous, competitive environments, ICBAC introduces a game-theoretic client selection mechanism using hedonic coalition formation. This enables supply chains to form stable, strategy-proof FL coalitions via preference-based selection without disclosing sensitive criteria. Extensive experiments on a Fabric testbed with a real-world dataset show ICBAC achieves blockchain performance comparable to static frameworks and provides effective anomaly detection under IID and non-IID data with zero raw-data sharing. ICBAC thus offers a practical, scalable solution for dynamic, privacy-preserving access control in decentralized supply chains.
Authors: Dong Pan, Bingtao Li, Yongsheng Zheng, Jiren Ma, Victor Fei
Abstract: The sparse Mixture of Experts(MoE) architecture has evolved as a powerful approach for scaling deep learning models to more parameters with comparable computation cost. As an important branch of large language model(LLM), MoE model only activate a subset of experts based on a routing network. This sparse conditional computation mechanism significantly improves computational efficiency, paving a promising path for greater scalability and cost-efficiency. It not only enhance downstream applications such as natural language processing, computer vision, and multimodal in various horizontal domains, but also exhibit broad applicability across vertical domains. Despite the growing popularity and application of MoE models across various domains, there lacks a systematic exploration of recent advancements of MoE in many important fields. Existing surveys on MoE suffer from limitations such as lack coverage or none extensively exploration of key areas. This survey seeks to fill these gaps. In this paper, Firstly, we examine the foundational principles of MoE, with an in-depth exploration of its core components-the routing network and expert network. Subsequently, we extend beyond the centralized paradigm to the decentralized paradigm, which unlocks the immense untapped potential of decentralized infrastructure, enables democratization of MoE development for broader communities, and delivers greater scalability and cost-efficiency. Furthermore we focus on exploring its vertical domain applications. Finally, we also identify key challenges and promising future research directions. To the best of our knowledge, this survey is currently the most comprehensive review in the field of MoE. We aim for this article to serve as a valuable resource for both researchers and practitioners, enabling them to navigate and stay up-to-date with the latest advancements.
Authors: Nanda Rani, Kimberly Milner, Minghao Shao, Meet Udeshi, Haoran Xi, Venkata Sai Charan Putrevu, Saksham Aggarwal, Sandeep K. Shukla, Prashanth Krishnamurthy, Farshad Khorrami, Muhammad Shafique, Ramesh Karri
Abstract: Real-world offensive security operations are inherently open-ended: attackers explore unknown attack surfaces, revise hypotheses under uncertainty, and operate without guaranteed success. Existing LLM-based offensive agent evaluations rely on closed-world settings with predefined goals and binary success criteria. To address this gap, we introduce CyberExplorer, an evaluation suite with two core components: (1) an open-environment benchmark built on a virtual machine hosting 40 vulnerable web services derived from real-world CTF challenges, where agents autonomously perform reconnaissance, target selection, and exploitation without prior knowledge of vulnerability locations; and (2) a reactive multi-agent framework supporting dynamic exploration without predefined plans. CyberExplorer enables fine-grained evaluation beyond flag recovery, capturing interaction dynamics, coordination behavior, failure modes, and vulnerability discovery signals-bridging the gap between benchmarks and realistic multi-target attack scenarios.
Authors: Ziyang Fan, Keyu Chen, Ruilong Xing, Yulin Li, Li Jiang, Zhuotao Tian
Abstract: Although Video Large Language Models (VLLMs) have shown remarkable capabilities in video understanding, they are required to process high volumes of visual tokens, causing significant computational inefficiency. Existing VLLMs acceleration frameworks usually compress spatial and temporal redundancy independently, which overlooks the spatiotemporal relationships, thereby leading to suboptimal spatiotemporal compression. The highly correlated visual features are likely to change in spatial position, scale, orientation, and other attributes over time due to the dynamic nature of video. Building on this insight, we introduce FlashVID, a training-free inference acceleration framework for VLLMs. Specifically, FlashVID utilizes Attention and Diversity-based Token Selection (ADTS) to select the most representative tokens for basic video representation, then applies Tree-based Spatiotemporal Token Merging (TSTM) for fine-grained spatiotemporal redundancy elimination. Extensive experiments conducted on three representative VLLMs across five video understanding benchmarks demonstrate the effectiveness and generalization of our method. Notably, by retaining only 10% of visual tokens, FlashVID preserves 99.1% of the performance of LLaVA-OneVision. Consequently, FlashVID can serve as a training-free and plug-and-play module for extending long video frames, which enables a 10x increase in video frame input to Qwen2.5-VL, resulting in a relative improvement of 8.6% within the same computational budget. Code is available at https://github.com/Fanziyang-v/FlashVID.
Authors: Yixuan Ye, Xuanyu Lu, Yuxin Jiang, Yuchao Gu, Rui Zhao, Qiwei Liang, Jiachun Pan, Fengda Zhang, Weijia Wu, Alex Jinpeng Wang
Abstract: World models aim to understand, remember, and predict dynamic visual environments, yet a unified benchmark for evaluating their fundamental abilities remains lacking. To address this gap, we introduce MIND, the first open-domain closed-loop revisited benchmark for evaluating Memory consIstency and action coNtrol in worlD models. MIND contains 250 high-quality videos at 1080p and 24 FPS, including 100 (first-person) + 100 (third-person) video clips under a shared action space and 25 + 25 clips across varied action spaces covering eight diverse scenes. We design an efficient evaluation framework to measure two core abilities: memory consistency and action control, capturing temporal stability and contextual coherence across viewpoints. Furthermore, we design various action spaces, including different character movement speeds and camera rotation angles, to evaluate the action generalization capability across different action spaces under shared scenes. To facilitate future performance benchmarking on MIND, we introduce MIND-World, a novel interactive Video-to-World baseline. Extensive experiments demonstrate the completeness of MIND and reveal key challenges in current world models, including the difficulty of maintaining long-term memory consistency and generalizing across action spaces. Project page: https://csu-jpg.github.io/MIND.github.io/
Authors: Isaac Han, Sangyeon Park, Seungwon Oh, Donghu Kim, Hojoon Lee, Kyung-Joong Kim
Abstract: Deep neural networks trained on nonstationary data must balance stability (i.e., retaining prior knowledge) and plasticity (i.e., adapting to new tasks). Standard reinitialization methods, which reinitialize weights toward their original values, are widely used but difficult to tune: conservative reinitializations fail to restore plasticity, while aggressive ones erase useful knowledge. We propose FIRE, a principled reinitialization method that explicitly balances the stability-plasticity tradeoff. FIRE quantifies stability through Squared Frobenius Error (SFE), measuring proximity to past weights, and plasticity through Deviation from Isometry (DfI), reflecting weight isotropy. The reinitialization point is obtained by solving a constrained optimization problem, minimizing SFE subject to DfI being zero, which is efficiently approximated by Newton-Schulz iteration. FIRE is evaluated on continual visual learning (CIFAR-10 with ResNet-18), language modeling (OpenWebText with GPT-0.1B), and reinforcement learning (HumanoidBench with SAC and Atari games with DQN). Across all domains, FIRE consistently outperforms both naive training without intervention and standard reinitialization methods, demonstrating effective balancing of the stability-plasticity tradeoff.
Authors: Boyang Xia, Weiyou Tian, Qingnan Ren, Jiaqi Huang, Jie Xiao, Shuo Lu, Kai Wang, Lynn Ai, Eric Yang, Bill Shi
Abstract: Training large language model (LLM) agents for adversarial games is often driven by episodic objectives such as win rate. In long-horizon settings, however, payoffs are shaped by latent strategic externalities that evolve over time, so myopic optimization and variation-based regret analyses can become vacuous even when the dynamics are predictable. To solve this problem, we introduce Implicit Strategic Optimization (ISO), a prediction-aware framework in which each agent forecasts the current strategic context and uses it to update its policy online. ISO combines a Strategic Reward Model (SRM) that estimates the long-run strategic value of actions with iso-grpo, a context-conditioned optimistic learning rule. We prove sublinear contextual regret and equilibrium convergence guarantees whose dominant terms scale with the number of context mispredictions; when prediction errors are bounded, our bounds recover the static-game rates obtained when strategic externalities are known. Experiments in 6-player No-Limit Texas Hold'em and competitive Pokemon show consistent improvements in long-term return over strong LLM and RL baselines, and graceful degradation under controlled prediction noise.
Authors: Yiheng Gao, Qin Hua, Zizhong Chen
Abstract: Algorithm-Based Fault Tolerance (ABFT) is widely adopted to detect silent data corruptions (SDCs) in matrix multiplication, a cornerstone operation in deep learning systems. However, existing threshold determination methods face critical challenges: analytical bounds are overly conservative, while probabilistic approaches like A-ABFT yield thresholds $160$--$4200\times$ larger than actual rounding errors. We present V-ABFT, a variance-based adaptive threshold algorithm that achieves tighter error bounds by directly modeling the verification difference. By leveraging statistical variance estimation, V-ABFT reduces the threshold-to-actual-error ratio to approximately $7$--$20\times$ for FP32/FP64 and $48$--$158\times$ for BF16, representing a \textbf{6--48$\times$ improvement} over A-ABFT while maintaining zero false positive rate across BF16, FP16, FP32, and FP64 precisions. Furthermore, we demonstrate that for fused-kernel ABFT implementations that verify before output quantization, low-precision GEMM can use FP32-level thresholds ($e_{\max} \approx 10^{-6}$), enabling \textbf{$\sim$1000$\times$ finer detection granularity} compared to offline verification with low-precision output ($e_{\max} \approx 10^{-3}$). We reproduce A-ABFT's experimental setup and validate our implementation against the original paper's results. Our method requires only $O(n)$ complexity using max/min/mean statistics, compared to A-ABFT's $O(pn)$ for finding $p$ largest values. Extensive experiments on synthetic data and real model weights (LLaMA-7B, GPT-2, ViT) demonstrate V-ABFT's effectiveness across diverse distributions. V-ABFT is platform-agnostic and has been integrated into fault-tolerant GEMM implementations on both NPUs and GPUs.
Authors: Manan Tayal, Mumuksh Tayal
Abstract: Offline reinforcement learning (RL) provides a compelling paradigm for training autonomous systems without the risks of online exploration, particularly in safety-critical domains. However, jointly achieving strong safety and performance from fixed datasets remains challenging. Existing safe offline RL methods often rely on soft constraints that allow violations, introduce excessive conservatism, or struggle to balance safety, reward optimization, and adherence to the data distribution. To address this, we propose Epigraph-Guided Flow Matching (EpiFlow), a framework that formulates safe offline RL as a state-constrained optimal control problem to co-optimize safety and performance. We learn a feasibility value function derived from an epigraph reformulation of the optimal control problem, thereby avoiding the decoupled objectives or post-hoc filtering common in prior work. Policies are synthesized by reweighting the behavior distribution based on this epigraph value function and fitting a generative policy via flow matching, enabling efficient, distribution-consistent sampling. Across various safety-critical tasks, including Safety-Gymnasium benchmarks, EpiFlow achieves competitive returns with near-zero empirical safety violations, demonstrating the effectiveness of epigraph-guided policy synthesis.
Authors: Yufei Wang, Haixu Liu, Tianxiang Xu, Chuancheng Shi, Hongsheng Xing
Abstract: To tackle the automatic recognition of "concealed emotions" in videos, this paper proposes a multimodal weak-supervision framework and achieves state-of-the-art results on the iMiGUE tennis-interview dataset. First, YOLO 11x detects and crops human portraits frame-by-frame, and DINOv2-Base extracts visual features from the cropped regions. Next, by integrating Chain-of-Thought and Reflection prompting (CoT + Reflection), Gemini 2.5 Pro automatically generates pseudo-labels and reasoning texts that serve as weak supervision for downstream models. Subsequently, OpenPose produces 137-dimensional key-point sequences, augmented with inter-frame offset features; the usual graph neural network backbone is simplified to an MLP to efficiently model the spatiotemporal relationships of the three key-point streams. An ultra-long-sequence Transformer independently encodes both the image and key-point sequences, and their representations are concatenated with BERT-encoded interview transcripts. Each modality is first pre-trained in isolation, then fine-tuned jointly, with pseudo-labeled samples merged into the training set for further gains. Experiments demonstrate that, despite severe class imbalance, the proposed approach lifts accuracy from under 0.6 in prior work to over 0.69, establishing a new public benchmark. The study also validates that an "MLP-ified" key-point backbone can match - or even surpass - GCN-based counterparts in this task.
Authors: Xihang Yu, Rajat Talak, Lorenzo Shaikewitz, Luca Carlone
Abstract: In the presence of occlusions and measurement noise, geometrically accurate scene reconstructions -- which fit the sensor data -- can still be physically incorrect. For instance, when estimating the poses and shapes of objects in the scene and importing the resulting estimates into a simulator, small errors might translate to implausible configurations including object interpenetration or unstable equilibrium. This makes it difficult to predict the dynamic behavior of the scene using a digital twin, an important step in simulation-based planning and control of contact-rich behaviors. In this paper, we posit that object pose and shape estimation requires reasoning holistically over the scene (instead of reasoning about each object in isolation), accounting for object interactions and physical plausibility. Towards this goal, our first contribution is Picasso, a physics-constrained reconstruction pipeline that builds multi-object scene reconstructions by considering geometry, non-penetration, and physics. Picasso relies on a fast rejection sampling method that reasons over multi-object interactions, leveraging an inferred object contact graph to guide samples. Second, we propose the Picasso dataset, a collection of 10 contact-rich real-world scenes with ground truth annotations, as well as a metric to quantify physical plausibility, which we open-source as part of our benchmark. Finally, we provide an extensive evaluation of Picasso on our newly introduced dataset and on the YCB-V dataset, and show it largely outperforms the state of the art while providing reconstructions that are both physically plausible and more aligned with human intuition.
Authors: Tong Zhang, Ru Zhang, Jianyi Liu
Abstract: The recent proliferation of diffusion models has made style mimicry effortless, enabling users to imitate unique artistic styles without authorization. In deployed platforms, this raises copyright and intellectual-property risks and calls for reliable protection. However, existing countermeasures either require costly weight editing as new styles emerge or rely on an explicitly specified editing style, limiting their practicality for deployment-side safety. To address this challenge, we propose DICE (Disentanglement of artist Style from Content via Contrastive Subspace Decomposition), a training-free framework for on-the-fly artist style erasure. Unlike style editing that require an explicitly specified replacement style, DICE performs style purification, removing the artist's characteristics while preserving the user-intended content. Our core insight is that a model cannot truly comprehend the artist style from a single text or image alone. Consequently, we abandon the traditional paradigm of identifying style from isolated samples. Instead, we construct contrastive triplets to compel the model to distinguish between style and non-style features in the latent space. By formalizing this disentanglement process as a solvable generalized eigenvalue problem, we achieve precise identification of the style subspace. Furthermore, we introduce an Adaptive Attention Decoupling Editing strategy dynamically assesses the style concentration of each token and performs differential suppression and content enhancement on the QKV vectors. Extensive experiments demonstrate that DICE achieves a superior balance between the thoroughness of style erasure and the preservation of content integrity. DICE introduces an additional overhead of only 3 seconds to disentangle style, providing a practical and efficient technique for curbing style mimicry.
Authors: Tianyu Li, Dongchen Han, Zixuan Cao, Haofeng Huang, Mengyu Zhou, Ming Chen, Erchao Zhao, Xiaoxi Jiang, Guanjun Jiang, Gao Huang
Abstract: Modern Transformers predominantly adopt the Pre-Norm paradigm for its optimization stability, foregoing the superior potential of the unstable Post-Norm architecture. Prior attempts to combine their strengths typically lead to a stability-performance trade-off. We attribute this phenomenon to a structural incompatibility within a single-stream design: Any application of the Post-Norm operation inevitably obstructs the clean identity gradient preserved by Pre-Norm. To fundamentally reconcile these paradigms, we propose SiameseNorm, a two-stream architecture that couples Pre-Norm-like and Post-Norm-like streams with shared parameters. This design decouples the optimization dynamics of the two streams, retaining the distinct characteristics of both Pre-Norm and Post-Norm by enabling all residual blocks to receive combined gradients inherited from both paradigms, where one stream secures stability while the other enhances expressivity. Extensive pre-training experiments on 1.3B-parameter models demonstrate that SiameseNorm exhibits exceptional optimization robustness and consistently outperforms strong baselines. Code is available at https://github.com/Qwen-Applications/SiameseNorm.
Authors: Sayantan Kumar, Peijie Qiu, Aristeidis Sotiras
Abstract: Normative modeling learns a healthy reference distribution and quantifies subject-specific deviations to capture heterogeneous disease effects. In Alzheimers disease (AD), multimodal neuroimaging offers complementary signals but VAE-based normative models often (i) fit the healthy reference distribution imperfectly, inflating false positives, and (ii) use posterior aggregation (e.g., PoE/MoE) that can yield weak multimodal fusion in the shared latent space. We propose mmSIVAE, a multimodal soft-introspective variational autoencoder combined with Mixture-of-Product-of-Experts (MOPOE) aggregation to improve reference fidelity and multimodal integration. We compute deviation scores in latent space and feature space as distances from the learned healthy distributions, and map statistically significant latent deviations to regional abnormalities for interpretability. On ADNI MRI regional volumes and amyloid PET SUVR, mmSIVAE improves reconstruction on held-out controls and produces more discriminative deviation scores for outlier detection than VAE baselines, with higher likelihood ratios and clearer separation between control and AD-spectrum cohorts. Deviation maps highlight region-level patterns aligned with established AD-related changes. More broadly, our results highlight the importance of training objectives that prioritize reference-distribution fidelity and robust multimodal posterior aggregation for normative modeling, with implications for deviation-based analysis across multimodal clinical data.
Authors: Valentin No\"el
Abstract: Deploying autonomous agents in the wild requires reliable safeguards against tool use failures. We propose a training free guardrail based on spectral analysis of attention topology that complements supervised approaches. On Llama 3.1 8B, our method achieves 97.7\% recall with multi-feature detection and 86.1\% recall with 81.0\% precision for balanced deployment, without requiring any labeled training data. Most remarkably, we discover that single layer spectral features act as near-perfect hallucination detectors: Llama L26 Smoothness achieves 98.2\% recall (213/217 hallucinations caught) with a single threshold, and Mistral L3 Entropy achieves 94.7\% recall. This suggests hallucination is not merely a wrong token but a thermodynamic state change: the model's attention becomes noise when it errs. Through controlled cross-model evaluation on matched domains ($N=1000$, $T=0.3$, same General domain, hallucination rates 20--22\%), we reveal the ``Loud Liar'' phenomenon: Llama 3.1 8B's failures are spectrally catastrophic and dramatically easier to detect, while Mistral 7B achieves the best discrimination (AUC 0.900). These findings establish spectral analysis as a principled, efficient framework for agent safety.
Authors: Shuyu Jiang, Hao Ren, Yichang Gao, Yi-Cheng Zhang, Li Qi, Dayong Xiao, Jie Fan, Rui Tang, Wei Wang
Abstract: Spreading dynamics is a central topic in the physics of complex systems and network science, providing a unified framework for understanding how information, behaviors, and diseases propagate through interactions among system units. In many propagation contexts, spreading processes are influenced by multiple interacting factors, such as information expression patterns, cultural contexts, living environments, cognitive preferences, and public policies, which are difficult to incorporate directly into classical modeling frameworks. Recently, large language models (LLMs) have exhibited strong capabilities in natural language understanding, reasoning, and generation, enabling explicit perception of semantic content and contextual cues in spreading processes, thereby supporting the analysis of the different influencing factors. Beyond serving as external analytical tools, LLMs can also act as interactive agents embedded in propagation systems, potentially influencing spreading pathways and feedback structures. Consequently, the roles and impacts of LLMs on spreading dynamics have become an active and rapidly growing research area across multiple research disciplines. This review provides a comprehensive overview of recent advances in applying LLMs to the study of spreading dynamics across two representative domains: digital epidemics, such as misinformation and rumors, and biological epidemics, including infectious disease outbreaks. We first examine the foundations of epidemic modeling from a complex-systems perspective and discuss how LLM-based approaches relate to traditional frameworks. We then systematically review recent studies from three key perspectives, which are epidemic modeling, epidemic detection and surveillance, and epidemic prediction and management, to clarify how LLMs enhance these areas. Finally, open challenges and potential research directions are discussed.
Authors: Mohammad Abu-Shaira, Weishi Shi
Abstract: LLMs are typically fine-tuned offline on domain-specific data, assuming a static domain. In practice, domain knowledge evolves continuously through new regulations, products, services, and interaction patterns. Retraining or fine-tuning LLMs for every new instance is computationally infeasible. Additionally, real-world environments also exhibit temporal dynamics with shifting data distributions. Disregarding this phenomenon, commonly referred to as concept drift, can significantly diminish a model's predictive accuracy. This mismatch between evolving domains and static adaptation pipelines highlights the need for efficient, real-time adaptation without costly retraining. In response, we introduce Online Domain-aware Decoding framework (ODD). ODD performs probability-level fusion between a base LLM and a prefix-tree prior, guided by adaptive confidence modulation using disagreement and continuity signals. Empirical evaluation under diverse drift scenarios demonstrates that ODD consistently surpasses LLM-Greedy and LLM-Temp Scaled across all syntactic and semantic NLG metrics. It yields an absolute ROUGE-L gain of 0.065 and a 13.6% relative improvement in Cosine Similarity over the best baseline. These results demonstrate ODD 's robustness to evolving lexical and contextual patterns, making it suitable for dynamic LLM applications.
Authors: Issar Tzachor, Dvir Samuel, Rami Ben-Ari
Abstract: Recent studies have adapted generative Multimodal Large Language Models (MLLMs) into embedding extractors for vision tasks, typically through fine-tuning to produce universal representations. However, their performance on video remains inferior to Video Foundation Models (VFMs). In this paper, we focus on leveraging MLLMs for video-text embedding and retrieval. We first conduct a systematic layer-wise analysis, showing that intermediate (pre-trained) MLLM layers already encode substantial task-relevant information. Leveraging this insight, we demonstrate that combining intermediate-layer embeddings with a calibrated MLLM head yields strong zero-shot retrieval performance without any training. Building on these findings, we introduce a lightweight text-based alignment strategy which maps dense video captions to short summaries and enables task-related video-text embedding learning without visual supervision. Remarkably, without any fine-tuning beyond text, our method outperforms current methods, often by a substantial margin, achieving state-of-the-art results across common video retrieval benchmarks.
Authors: Jasmine Cui, Charles Ye
Abstract: What happens when a language model thinks without words? Standard reasoning LLMs verbalize intermediate steps as chain-of-thought; latent reasoning transformers (LRTs) instead perform deliberation entirely in continuous hidden space. We investigate an LRT, decoding the model's evolving beliefs at every step on a multiple-choice QA benchmark. We find that the model spontaneously learns a structured search process in latent space. Deliberation follows a consistent trajectory: an exploration phase where probability mass spreads across candidates, tentative commitment to a frontrunner, and either convergence or backtracking. Backtracking is prevalent (32% of instances), beneficial (34% accuracy gain over non-backtracking instances), and predominantly directed away from the semantically closest distractor toward the correct answer. The search is adaptive: replacing distractors with implausible alternatives shortens exploration by 54%. Latent reasoning models achieve in activation space what chain-of-thought achieves through words: the ability to be wrong, notice, and recover.
Authors: Hoang Giang Pham, Tien Mai
Abstract: The mixed logit model is a flexible and widely used demand model in pricing and revenue management. However, existing work on mixed-logit pricing largely focuses on unconstrained settings, limiting its applicability in practice where prices are subject to business or regulatory constraints. We study the constrained pricing problem under multinomial and mixed logit demand models. For the multinomial logit model, corresponding to a single customer segment, we show that the constrained pricing problem admits a polynomial-time approximation scheme (PTAS) via a reformulation based on exponential cone programming, yielding an $\varepsilon$-optimal solution in polynomial time. For finite mixed logit models with $T$ customer segments, we reformulate the problem as a bilinear exponential cone program with $O(T)$ bilinear terms. This structure enables a Branch-and-Bound algorithm whose complexity is exponential only in $T$. Consequently, constrained pricing under finite mixtures of logit admits a PTAS when the number of customer segments is bounded. Numerical experiments demonstrate strong performance relative to state-of-the-art baselines.
Authors: Ke Xu, Shera Potka, Alex Thomo
Abstract: Large Language Models are increasingly employed in generating consumer product recommendations, yet their potential for embedding and amplifying gender and race biases remains underexplored. This paper serves as one of the first attempts to examine these biases within LLM-generated recommendations. We leverage prompt engineering to elicit product suggestions from LLMs for various race and gender groups and employ three analytical methods-Marked Words, Support Vector Machines, and Jensen-Shannon Divergence-to identify and quantify biases. Our findings reveal significant disparities in the recommendations for demographic groups, underscoring the need for more equitable LLM recommendation systems.
Authors: Md Rafi Ur Rashid, MD Sadik Hossain Shanto, Vishnu Asutosh Dasu, Shagufta Mehnaz
Abstract: Vision-Language Models (VLMs) are now a core part of modern AI. Recent work proposed several visual jailbreak attacks using single/ holistic images. However, contemporary VLMs demonstrate strong robustness against such attacks due to extensive safety alignment through preference optimization (e.g., RLHF). In this work, we identify a new vulnerability: while VLM pretraining and instruction tuning generalize well to split-image inputs, safety alignment is typically performed only on holistic images and does not account for harmful semantics distributed across multiple image fragments. Consequently, VLMs often fail to detect and refuse harmful split-image inputs, where unsafe cues emerge only after combining images. We introduce novel split-image visual jailbreak attacks (SIVA) that exploit this misalignment. Unlike prior optimization-based attacks, which exhibit poor black-box transferability due to architectural and prior mismatches across models, our attacks evolve in progressive phases from naive splitting to an adaptive white-box attack, culminating in a black-box transfer attack. Our strongest strategy leverages a novel adversarial knowledge distillation (Adv-KD) algorithm to substantially improve cross-model transferability. Evaluations on three state-of-the-art modern VLMs and three jailbreak datasets demonstrate that our strongest attack achieves up to 60% higher transfer success than existing baselines. Lastly, we propose efficient ways to address this critical vulnerability in the current VLM safety alignment.
Authors: Xinyu Yang, Junlin Han, Rishi Bommasani, Jinqi Luo, Wenjie Qu, Wangchunshu Zhou, Adel Bibi, Xiyao Wang, Jaehong Yoon, Elias Stengel-Eskin, Shengbang Tong, Lingfeng Shen, Rafael Rafailov, Runjia Li, Zhaoyang Wang, Yiyang Zhou, Chenhang Cui, Yu Wang, Wenhao Zheng, Huichi Zhou, Jindong Gu, Zhaorun Chen, Peng Xia, Tony Lee, Thomas Zollo, Vikash Sehwag, Jixuan Leng, Jiuhai Chen, Yuxin Wen, Huan Zhang, Zhun Deng, Linjun Zhang, Pavel Izmailov, Pang Wei Koh, Yulia Tsvetkov, Andrew Wilson, Jiaheng Zhang, James Zou, Cihang Xie, Hao Wang, Philip Torr, Julian McAuley, David Alvarez-Melis, Florian Tram\`er, Kaidi Xu, Suman Jana, Chris Callison-Burch, Rene Vidal, Filippos Kokkinos, Mohit Bansal, Beidi Chen, Huaxiu Yao
Abstract: Foundation models, including Large Language Models (LLMs), Multimodal Large Language Models (MLLMs), Image Generative Models (i.e, Text-to-Image Models and Image-Editing Models), and Video Generative Models, have become essential tools with broad applications across various domains such as law, medicine, education, finance, science, and beyond. As these models see increasing real-world deployment, ensuring their reliability and responsibility has become critical for academia, industry, and government. This survey addresses the reliable and responsible development of foundation models. We explore critical issues, including bias and fairness, security and privacy, uncertainty, explainability, and distribution shift. Our research also covers model limitations, such as hallucinations, as well as methods like alignment and Artificial Intelligence-Generated Content (AIGC) detection. For each area, we review the current state of the field and outline concrete future research directions. Additionally, we discuss the intersections between these areas, highlighting their connections and shared challenges. We hope our survey fosters the development of foundation models that are not only powerful but also ethical, trustworthy, reliable, and socially responsible.
Authors: Sahana Ramnath, Nima Chitsazan, Mingyang Zhou, Chia-Hsuan Lee, Shi-Xiong Zhang, Stephen Rawls, Sambit Sahu, Sangwoo Cho, Xiang Ren, Genta Indra Winata, Akshaj Kumar Veldanda
Abstract: Dialogues are a predominant mode of communication for humans, and it is immensely helpful to have automatically generated summaries of them (e.g., to revise key points discussed in a meeting, to review conversations between customer agents and product users). Prior works on dialogue summary evaluation largely ignore the complexities specific to this task: (i) shift in structure, from multiple speakers discussing information in a scattered fashion across several turns, to a summary's sentences, and (ii) shift in narration viewpoint, from speakers' first/second-person narration, standardized third-person narration in the summary. In this work, we introduce our framework DIALSUMMER to address the above. We propose DIAL-SUMMER's taxonomy of errors to comprehensively evaluate dialogue summaries at two hierarchical levels: DIALOGUE-LEVEL that focuses on the broader speakers/turns, and WITHIN-TURN-LEVEL that focuses on the information talked about inside a turn. We then present DIAL-SUMMER's dataset composed of dialogue summaries manually annotated with our taxonomy's fine-grained errors. We conduct empirical analyses of these annotated errors, and observe interesting trends (e.g., turns occurring in middle of the dialogue are the most frequently missed in the summary, extrinsic hallucinations largely occur at the end of the summary). We also conduct experiments on LLM-Judges' capability at detecting these errors, through which we demonstrate the challenging nature of our dataset, the robustness of our taxonomy, and the need for future work in this field to enhance LLMs' performance in the same. Code and inference dataset coming soon.
Authors: Seonglae Cho, Zekun Wu, Kleyton Da Costa, Adriano Koshiyama
Abstract: When a language model asserts that "the capital of Australia is Sydney," does it know this is wrong? We characterize the geometry of correctness representations across 9 models from 5 architecture families. The structure is simple: the discriminative signal occupies 3-8 dimensions, performance degrades with additional dimensions, and no nonlinear classifier improves over linear separation. Centroid distance in the low-dimensional subspace matches trained probe performance (0.90 AUC), enabling few-shot detection: on GPT-2, 25 labeled examples achieve 89% of full-data accuracy. We validate causally through activation steering: the learned direction produces 10.9 percentage point changes in error rates while random directions show no effect. Internal probes achieve 0.80-0.97 AUC; output-based methods (P(True), semantic entropy) achieve only 0.44-0.64 AUC. The correctness signal exists internally but is not expressed in outputs. That centroid distance matches probe performance indicates class separation is a mean shift, making detection geometric rather than learned.
Authors: Milan Ganai, Katie Luo, Jonas Frey, Clark Barrett, Marco Pavone
Abstract: Embodied Chain-of-Thought (CoT) reasoning has significantly enhanced Vision-Language-Action (VLA) models, yet current methods rely on rigid templates to specify reasoning primitives (e.g., objects in the scene, high-level plans, structural affordances). These templates can force policies to process irrelevant information that distracts from critical action-prediction signals. This creates a bottleneck: without successful policies, we cannot verify reasoning quality; without quality reasoning, we cannot build robust policies. We introduce R&B-EnCoRe, which enables models to bootstrap embodied reasoning from internet-scale knowledge through self-supervised refinement. By treating reasoning as a latent variable within importance-weighted variational inference, models can generate and distill a refined reasoning training dataset of embodiment-specific strategies without external rewards, verifiers, or human annotation. We validate R&B-EnCoRe across manipulation (Franka Panda in simulation, WidowX in hardware), legged navigation (bipedal, wheeled, bicycle, quadruped), and autonomous driving embodiments using various VLA architectures with 1B, 4B, 7B, and 30B parameters. Our approach achieves 28% gains in manipulation success, 101% improvement in navigation scores, and 21% reduction in collision-rate metric over models that indiscriminately reason about all available primitives. R&B-EnCoRe enables models to distill reasoning that is predictive of successful control, bypassing manual annotation engineering while grounding internet-scale knowledge in physical execution.
Authors: Tianji Cong, Yuanyuan Tian, Andreas Mueller, Rathijit Sen, Yeye He, Fotis Psallidas, Shaleen Deep, H. V. Jagadish
Abstract: Automatically inferring join relationships is a critical task for effective data discovery, integration, querying and reuse. However, accurately and efficiently identifying these relationships in large and complex schemas can be challenging, especially in enterprise settings where access to data values is constrained. In this paper, we introduce the problem of join graph inference when only metadata is available. We conduct an empirical study on a large number of real-world schemas and observe that join graphs when represented as adjacency matrices exhibit two key properties: high sparsity and low-rank structure. Based on these novel observations, we formulate join graph inference as a low-rank matrix completion problem and propose Nexus, an end-to-end solution using only metadata. To further enhance accuracy, we propose a novel Expectation-Maximization algorithm that alternates between low-rank matrix completion and refining join candidate probabilities by leveraging Large Language Models. Our extensive experiments demonstrate that Nexus outperforms existing methods by a significant margin on four datasets including a real-world production dataset. Additionally, Nexus can operate in a fast mode, providing comparable results with up to 6x speedup, offering a practical and efficient solution for real-world deployments.
Authors: Cindy Peng, Megan Chai, Gao Mo, Naveen Raman, Ningjing Tang, Shannon Pagdon, Margaret Swarbrick, Nev Jones, Fei Fang, Hong Shen
Abstract: Peer-run organizations (PROs) provide critical, recovery-based behavioral health support rooted in lived experience. As large language models (LLMs) enter this domain, their scale, conversationality, and opacity introduce new challenges for situatedness, trust, and autonomy. Partnering with Collaborative Support Programs of New Jersey (CSPNJ), a statewide PRO in the Northeastern United States, we used comicboarding, a co-design method, to conduct workshops with 16 peer specialists and 10 service users exploring perceptions of integrating an LLM-based recommendation system into peer support. Findings show that depending on how LLMs are introduced, constrained, and co-used, they can reconfigure in-room dynamics by sustaining, undermining, or amplifying the relational authority that grounds peer support. We identify opportunities, risks, and mitigation strategies across three tensions: bridging scale and locality, protecting trust and relational dynamics, and preserving peer autonomy amid efficiency gains. We contribute design implications that center lived-experience-in-the-loop, reframe trust as co-constructed, and position LLMs not as clinical tools but as relational collaborators in high-stakes, community-led care.
Authors: Konstantinos Mitsides, Maxence Faldor, Antoine Cully
Abstract: Open-ended learning frames intelligence as emerging from continual interaction with an ever-expanding space of environments. While recent advances have utilized foundation models to programmatically generate diverse environments, these approaches often focus on discovering isolated behaviors rather than orchestrating sustained progression. In complex open-ended worlds, the large combinatorial space of possible challenges makes it difficult for agents to discover sequences of experiences that remain consistently learnable. To address this, we propose Dreaming in Code (DiCode), a framework in which foundation models synthesize executable environment code to scaffold learning toward increasing competence. In DiCode, "dreaming" takes the form of materializing code-level variations of the world. We instantiate DiCode in Craftax, a challenging open-ended benchmark characterized by rich mechanics and long-horizon progression. Empirically, DiCode enables agents to acquire long-horizon skills, achieving a $16\%$ improvement in mean return over the strongest baseline and non-zero success on late-game combat tasks where prior methods fail. Our results suggest that code-level environment design provides a practical mechanism for curriculum control, enabling the construction of intermediate environments that bridge competence gaps in open-ended worlds. Project page and source code are available at https://konstantinosmitsides.github.io/dreaming-in-code and https://github.com/konstantinosmitsides/dreaming-in-code.
URLs: https://konstantinosmitsides.github.io/dreaming-in-code, https://github.com/konstantinosmitsides/dreaming-in-code.
Authors: Haoran Liu, Zheni Zeng, Yukun Yan, Yuxuan Chen, Yunduo Xiao
Abstract: Molecule generation and optimization is a fundamental task in chemical domain. The rapid development of intelligent tools, especially large language models (LLMs) with powerful knowledge reserves and interactive capabilities, has provided new paradigms for it. Nevertheless, the intrinsic challenge for LLMs lies in the complex implicit relationship between molecular structure and pharmacological properties and the lack of corresponding labeled data. To bridge this gap, we propose DrugR, an LLM-based method that introduces explicit, step-by-step pharmacological reasoning into the optimization process. Our approach integrates domain-specific continual pretraining, supervised fine-tuning via reverse data engineering, and self-balanced multi-granular reinforcement learning. This framework enables DrugR to effectively improve key ADMET properties while preserving the original molecule's core efficacy. Experimental results demonstrate that DrugR achieves comprehensive enhancement across multiple properties without compromising structural similarity or target binding affinity. Importantly, its explicit reasoning process provides clear, interpretable rationales for each optimization step, yielding actionable design insights and advancing toward automated, knowledge-driven scientific discovery. Our code and model checkpoints are open-sourced to foster future research.
Authors: Huan Zhang, Yanjian Zhang, Guillaume Wisniewski, Nadi Tomeh, Bang Liu
Abstract: We propose a sparsity-aware evolutionary (SAE) framework for model merging that involves iterative pruning-merging cycles to act as a novel mutation operator. We incorporate the sparsity constraints into the score function, which steers the evolutionary process to favor more sparse models, in addition to other conventional performance scores. Interestingly, the by-product of \textit{competition} for sparsity introduces an extra local \textit{attraction} and interplay into the evolutionary process: if one competitor has more zero elements, the other competitor's non-zero elements will occupy those positions, even though the less sparse competitor loses to the more sparse competitor in other positions. The proposed pipeline is evaluated on a variety of large-scale LLM benchmarks. Experiments demonstrate that our approach can improve model merging reliability across multiple benchmarks, and is easy to incorporate due to its simplicity and being orthogonal to most existing approaches.
Authors: Xuhua Ma, Richong Zhang, Zhijie Nie
Abstract: Retrieval-Augmented Generation (RAG) often struggles with knowledge conflicts, where model-internal parametric knowledge overrides retrieved evidence, leading to unfaithful outputs. Existing approaches are often limited, relying either on superficial decoding adjustments or weight editing that necessitates ground-truth targets. Through layer-wise analysis, we attribute this failure to a parametric suppression phenomenon: specifically, in deep layers, certain FFN layers overwrite context-sensitive representations with memorized priors. To address this, we propose CoRect (Context-Aware Logit Contrast for Hidden State Rectification). By contrasting logits from contextualized and non-contextualized forward passes, CoRect identifies layers that exhibit high parametric bias without requiring ground-truth labels. It then rectifies the hidden states to preserve evidence-grounded information. Across question answering (QA) and summarization benchmarks, CoRect consistently improves faithfulness and reduces hallucinations compared to strong baselines.
Authors: Rama Adithya, Varanasi, Nov, Oded, Wiesenfeld, Batia Mishan
Abstract: This study investigates how professional writers' complex relationship with GenAI shapes their work practices and outcomes. Through a cross-sectional survey with writing professionals (n=403) in diverse roles, we show that collaboration and rivalry orientation are associated with differences in work practices and outcomes. Rivalry is primarily associated with relational crafting and skill maintenance. Collaboration is primarily associated with task crafting, productivity, and satisfaction, at the cost of long-term skill deterioration. Combination of the orientations (high rivalry and high collaboration) reconciles these differences, while boosting the association with the outcomes. Our findings argue for a balanced approach where high levels of rivalry and collaboration are essential to shape work practices and generate outcomes aimed at the long-term success of the job. We present key design implications on how to increase friction (rivalry) and reduce over-reliance (collaboration) to achieve a more balanced relationship with GenAI.
Authors: Hongwei Ren, Youxin Jiang, Qifei Gu, Xiangqian Wu
Abstract: Event cameras have been widely adopted in safety-critical domains such as autonomous driving, robotics, and human-computer interaction. A pressing challenge arises from the vulnerability of deep neural networks to adversarial examples, which poses a significant threat to the reliability of event-based systems. Nevertheless, research into adversarial attacks on events is scarce. This is primarily due to the non-differentiable nature of mainstream event representations, which hinders the extension of gradient-based attack methods. In this paper, we propose MA-ADV, a novel \textbf{M}otion-\textbf{A}ware \textbf{Adv}ersarial framework. To the best of our knowledge, this is the first work to generate adversarial events by leveraging point cloud representations. MA-ADV accounts for high-frequency noise in events and employs a diffusion-based approach to smooth perturbations, while fully leveraging the spatial and temporal relationships among events. Finally, MA-ADV identifies the minimal-cost perturbation through a combination of sample-wise Adam optimization, iterative refinement, and binary search. Extensive experimental results validate that MA-ADV ensures a 100\% attack success rate with minimal perturbation cost, and also demonstrate enhanced robustness against defenses, underscoring the critical security challenges facing future event-based perception systems.
Authors: Jiatao Chen, Xing Tang, Xiaoyue Duan, Yutang Feng, Jinchao Zhang, Jie Zhou
Abstract: While existing Singing Voice Synthesis systems achieve high-fidelity solo performances, they are constrained by global timbre control, failing to address dynamic multi-singer arrangement and vocal texture within a single song. To address this, we propose Tutti, a unified framework designed for structured multi-singer generation. Specifically, we introduce a Structure-Aware Singer Prompt to enable flexible singer scheduling evolving with musical structure, and propose Complementary Texture Learning via Condition-Guided VAE to capture implicit acoustic textures (e.g., spatial reverberation and spectral fusion) that are complementary to explicit controls. Experiments demonstrate that Tutti excels in precise multi-singer scheduling and significantly enhances the acoustic realism of choral generation, offering a novel paradigm for complex multi-singer arrangement. Audio samples are available at https://annoauth123-ctrl.github.io/Tutii_Demo/.
Authors: Jaylen Jones, Zhehao Zhang, Yuting Ning, Eric Fosler-Lussier, Pierre-Luc St-Charles, Yoshua Bengio, Dawn Song, Yu Su, Huan Sun
Abstract: Although computer-use agents (CUAs) hold significant potential to automate increasingly complex OS workflows, they can demonstrate unsafe unintended behaviors that deviate from expected outcomes even under benign input contexts. However, exploration of this risk remains largely anecdotal, lacking concrete characterization and automated methods to proactively surface long-tail unintended behaviors under realistic CUA scenarios. To fill this gap, we introduce the first conceptual and methodological framework for unintended CUA behaviors, by defining their key characteristics, automatically eliciting them, and analyzing how they arise from benign inputs. We propose AutoElicit: an agentic framework that iteratively perturbs benign instructions using CUA execution feedback, and elicits severe harms while keeping perturbations realistic and benign. Using AutoElicit, we surface hundreds of harmful unintended behaviors from state-of-the-art CUAs such as Claude 4.5 Haiku and Opus. We further evaluate the transferability of human-verified successful perturbations, identifying persistent susceptibility to unintended behaviors across various other frontier CUAs. This work establishes a foundation for systematically analyzing unintended behaviors in realistic computer-use settings.
Authors: Shoubin Yu, Yue Zhang, Zun Wang, Jaehong Yoon, Huaxiu Yao, Mingyu Ding, Mohit Bansal
Abstract: Despite rapid progress in Multimodal Large Language Models (MLLMs), visual spatial reasoning remains unreliable when correct answers depend on how a scene would appear under unseen or alternative viewpoints. Recent work addresses this by augmenting reasoning with world models for visual imagination, but questions such as when imagination is actually necessary, how much of it is beneficial, and when it becomes harmful, remain poorly understood. In practice, indiscriminate imagination can increase computation and even degrade performance by introducing misleading evidence. In this work, we present an in-depth analysis of test-time visual imagination as a controllable resource for spatial reasoning. We study when static visual evidence is sufficient, when imagination improves reasoning, and how excessive or unnecessary imagination affects accuracy and efficiency. To support this analysis, we introduce AVIC, an adaptive test-time framework with world models that explicitly reasons about the sufficiency of current visual evidence before selectively invoking and scaling visual imagination. Across spatial reasoning benchmarks (SAT, MMSI) and an embodied navigation benchmark (R2R), our results reveal clear scenarios where imagination is critical, marginal, or detrimental, and show that selective control can match or outperform fixed imagination strategies with substantially fewer world-model calls and language tokens. Overall, our findings highlight the importance of analyzing and controlling test-time imagination for efficient and reliable spatial reasoning.
Authors: Zahra Rahimi Afzal, Tara Esmaeilbeig, Mojtaba Soltanalian, Mesrob I. Ohannessian
Abstract: Parameter-Efficient Fine-Tuning (PEFT) is a popular class of techniques that strive to adapt large models in a scalable and resource-efficient manner. Yet, the mechanisms underlying their training performance and generalization remain underexplored. In this paper, we provide several insights into such fine-tuning through the lens of linearization. Fine-tuned models are often implicitly encouraged to remain close to the pretrained model. By making this explicit, using an Euclidean distance inductive bias in parameter space, we show that fine-tuning dynamics become equivalent to learning with the positive-definite neural tangent kernel (NTK). We specifically analyze how close the fully linear and the linearized fine-tuning optimizations are, based on the strength of the regularization. This allows us to be pragmatic about how good a model linearization is when fine-tuning large language models (LLMs). When linearization is a good model, our findings reveal a strong correlation between the eigenvalue spectrum of the NTK and the performance of model adaptation. Motivated by this, we give spectral perturbation bounds on the NTK induced by the choice of layers selected for fine-tuning. We empirically validate our theory on Low Rank Adaptation (LoRA) on LLMs. These insights not only characterize fine-tuning but also have the potential to enhance PEFT techniques, paving the way to better informed and more nimble adaptation in LLMs.
Authors: Juncheng Dong, Bowen He, Moyang Guo, Ethan X. Fang, Zhuoran Yang, Vahid Tarokh
Abstract: In-context reinforcement learning (ICRL) leverages the in-context learning capabilities of transformer models (TMs) to efficiently generalize to unseen sequential decision-making tasks without parameter updates. However, existing ICRL methods rely on explicit reward signals during pretraining, which limits their applicability when rewards are ambiguous, hard to specify, or costly to obtain. To overcome this limitation, we propose a new learning paradigm, In-Context Preference-based Reinforcement Learning (ICPRL), in which both pretraining and deployment rely solely on preference feedback, eliminating the need for reward supervision. We study two variants that differ in the granularity of feedback: Immediate Preference-based RL (I-PRL) with per-step preferences, and Trajectory Preference-based RL (T-PRL) with trajectory-level comparisons. We first show that supervised pretraining, a standard approach in ICRL, remains effective under preference-only context datasets, demonstrating the feasibility of in-context reinforcement learning using only preference signals. To further improve data efficiency, we introduce alternative preference-native frameworks for I-PRL and T-PRL that directly optimize TM policies from preference data without requiring reward signals nor optimal action labels.Experiments on dueling bandits, navigation, and continuous control tasks demonstrate that ICPRL enables strong in-context generalization to unseen tasks, achieving performance comparable to ICRL methods trained with full reward supervision.
Authors: Jinhao Li, Yuxuan Cong, Yingqiao Wang, Hao Xia, Shan Huang, Yijia Zhang, Ningyi Xu, Guohao Dai
Abstract: Diffusion policies have recently emerged as a powerful paradigm for visuomotor control in robotic manipulation due to their ability to model the distribution of action sequences and capture multimodality. However, iterative denoising leads to substantial inference latency, limiting control frequency in real-time closed-loop systems. Existing acceleration methods either reduce sampling steps, bypass diffusion through direct prediction, or reuse past actions, but often struggle to jointly preserve action quality and achieve consistently low latency. In this work, we propose STEP, a lightweight spatiotemporal consistency prediction mechanism to construct high-quality warm-start actions that are both distributionally close to the target action and temporally consistent, without compromising the generative capability of the original diffusion policy. Then, we propose a velocity-aware perturbation injection mechanism that adaptively modulates actuation excitation based on temporal action variation to prevent execution stall especially for real-world tasks. We further provide a theoretical analysis showing that the proposed prediction induces a locally contractive mapping, ensuring convergence of action errors during diffusion refinement. We conduct extensive evaluations on nine simulated benchmarks and two real-world tasks. Notably, STEP with 2 steps can achieve an average 21.6% and 27.5% higher success rate than BRIDGER and DDIM on the RoboMimic benchmark and real-world tasks, respectively. These results demonstrate that STEP consistently advances the Pareto frontier of inference latency and success rate over existing methods.
Authors: Jinwoo Kim, S\'ekou-Oumar Kaba, Jiyun Park, Seunghoon Hong, Siamak Ravanbakhsh
Abstract: We study the problem of transformation inversion on general Lie groups: a datum is transformed by an unknown group element, and the goal is to recover an inverse transformation that maps it back to the original data distribution. Such unknown transformations arise widely in machine learning and scientific modeling, where they can significantly distort observations. We take a probabilistic view and model the posterior over transformations as a Boltzmann distribution defined by an energy function on data space. To sample from this posterior, we introduce a diffusion process on Lie groups that keeps all updates on-manifold and only requires computations in the associated Lie algebra. Our method, Transformation-Inverting Energy Diffusion (TIED), relies on a new trivialized target-score identity that enables efficient score-based sampling of the transformation posterior. As a key application, we focus on test-time equivariance, where the objective is to improve the robustness of pretrained neural networks to input transformations. Experiments on image homographies and PDE symmetries demonstrate that TIED can restore transformed inputs to the training distribution at test time, showing improved performance over strong canonicalization and sampling baselines. Code is available at https://github.com/jw9730/tied.
Authors: Junwei Su, Chuan Wu
Abstract: Reinforcement Learning (RL) has emerged as a crucial method for training or fine-tuning large language models (LLMs), enabling adaptive, task-specific optimizations through interactive feedback. Multi-Agent Reinforcement Learning (MARL), in particular, offers a promising avenue by decomposing complex tasks into specialized subtasks learned by distinct interacting agents, potentially enhancing the ability and efficiency of LLM systems. However, theoretical insights regarding when and why MARL outperforms Single-Agent RL (SARL) remain limited, creating uncertainty in selecting the appropriate RL framework. In this paper, we address this critical gap by rigorously analyzing the comparative sample efficiency of MARL and SARL within the context of LLM. Leveraging the Probably Approximately Correct (PAC) framework, we formally define SARL and MARL setups for LLMs, derive explicit sample complexity bounds, and systematically characterize how task decomposition and alignment influence learning efficiency. Our results demonstrate that MARL improves sample complexity when tasks naturally decompose into independent subtasks, whereas dependent subtasks diminish MARL's comparative advantage. Additionally, we introduce and analyze the concept of task alignment, quantifying the trade-offs when enforcing independent task decomposition despite potential misalignments. These theoretical insights clarify empirical inconsistencies and provide practical criteria for deploying MARL strategies effectively in complex LLM scenarios.
Authors: Jan Philip Wahle
Abstract: Language enables humans to share knowledge, reason about the world, and pass on strategies for survival and innovation across generations. At the heart of this process is not just the ability to communicate but also the remarkable flexibility in how we can express ourselves. We can express the same thoughts in virtually infinite ways using different words and structures - this ability to rephrase and reformulate expressions is known as paraphrase. Modeling paraphrases is a keystone to meaning in computational language models; being able to construct different variations of texts that convey the same meaning or not shows strong abilities of semantic understanding. If computational language models are to represent meaning, they must understand and control the different aspects that construct the same meaning as opposed to different meanings at a fine granularity. Yet most existing approaches reduce paraphrasing to a binary decision between two texts or to producing a single rewrite of a source, obscuring which linguistic factors are responsible for meaning preservation. In this thesis, I propose that decomposing paraphrases into their constituent linguistic aspects (paraphrase types) offers a more fine-grained and cognitively grounded view of semantic equivalence. I show that even advanced machine learning models struggle with this task. Yet, when explicitly trained on paraphrase types, models achieve stronger performance on related paraphrase tasks and downstream applications. For example, in plagiarism detection, language models trained on paraphrase types surpass human baselines: 89.6% accuracy compared to 78.4% for plagiarism cases from Wikipedia, and 66.5% compared to 55.7% for plagiarism of scientific papers from arXiv. In identifying duplicate questions on Quora, models trained with paraphrase types improve over models trained on binary pairs. Furthermore, I demonstrate that...
Authors: Xiangbo Gao, Renjie Li, Xinghao Chen, Yuheng Wu, Suofei Feng, Qing Yin, Zhengzhong Tu
Abstract: The landscape of AI video generation is undergoing a pivotal shift: moving beyond general generation - which relies on exhaustive prompt-engineering and "cherry-picking" - towards fine-grained, controllable generation and high-fidelity post-processing. In professional AI-assisted filmmaking, it is crucial to perform precise, targeted modifications. A cornerstone of this transition is video instance insertion, which requires inserting a specific instance into existing footage while maintaining scene integrity. Unlike traditional video editing, this task demands several requirements: precise spatial-temporal placement, physically consistent scene interaction, and the faithful preservation of original dynamics - all achieved under minimal user effort. In this paper, we propose PISCO, a video diffusion model for precise video instance insertion with arbitrary sparse keyframe control. PISCO allows users to specify a single keyframe, start-and-end keyframes, or sparse keyframes at arbitrary timestamps, and automatically propagates object appearance, motion, and interaction. To address the severe distribution shift induced by sparse conditioning in pretrained video diffusion models, we introduce Variable-Information Guidance for robust conditioning and Distribution-Preserving Temporal Masking to stabilize temporal generation, together with geometry-aware conditioning for realistic scene adaptation. We further construct PISCO-Bench, a benchmark with verified instance annotations and paired clean background videos, and evaluate performance using both reference-based and reference-free perceptual metrics. Experiments demonstrate that PISCO consistently outperforms strong inpainting and video editing baselines under sparse control, and exhibits clear, monotonic performance improvements as additional control signals are provided. Project page: xiangbogaobarry.github.io/PISCO.
Authors: Haixu Liu, Yufei Wang, Tianxiang Xu, Chuancheng Shi, Hongsheng Xing
Abstract: Large-scale, cross-species plant distribution prediction plays a crucial role in biodiversity conservation, yet modeling efforts in this area still face significant challenges due to the sparsity and bias of observational data. Presence-Absence (PA) data provide accurate and noise-free labels, but are costly to obtain and limited in quantity; Presence-Only (PO) data, by contrast, offer broad spatial coverage and rich spatiotemporal distribution, but suffer from severe label noise in negative samples. To address these real-world constraints, this paper proposes a multimodal fusion framework that fully leverages the strengths of both PA and PO data. We introduce an innovative pseudo-label aggregation strategy for PO data based on the geographic coverage of satellite imagery, enabling geographic alignment between the label space and remote sensing feature space. In terms of model architecture, we adopt Swin Transformer Base as the backbone for satellite imagery, utilize the TabM network for tabular feature extraction, retain the Temporal Swin Transformer for time-series modeling, and employ a stackable serial tri-modal cross-attention mechanism to optimize the fusion of heterogeneous modalities. Furthermore, empirical analysis reveals significant geographic distribution shifts between PA training and test samples, and models trained by directly mixing PO and PA data tend to experience performance degradation due to label noise in PO data. To address this, we draw on the mixture-of-experts paradigm: test samples are partitioned according to their spatial proximity to PA samples, and different models trained on distinct datasets are used for inference and post-processing within each partition. Experiments on the GeoLifeCLEF 2025 dataset demonstrate that our approach achieves superior predictive performance in scenarios with limited PA coverage and pronounced distribution shifts.
Authors: Themistoklis Haris, Zihan Zhang, Yuichi Yoshida
Abstract: Understanding simplicity biases in deep learning offers a promising path toward developing reliable AI. A common metric for this, inspired by Boolean function analysis, is average sensitivity, which captures a model's robustness to single-token perturbations. We argue that average sensitivity has two key limitations: it lacks a natural generalization to real-valued domains and fails to explain the "junta-like" input dependence we empirically observe in modern LLMs. To address these limitations, we propose noise stability as a more comprehensive simplicity metric. Noise stability expresses a model's robustness to correlated noise applied to all input coordinates simultaneously. We provide a theoretical analysis of noise stability for single-layer attention and ReLU MLP layers and tackle the multi-layer propagation problem with a covariance interval propagation approach. Building on this theory, we develop a practical noise stability regularization method. Experiments on algorithmic and next-token-prediction tasks show that our regularizer consistently catalyzes grokking and accelerates training by approximately $35\%$ and $75\%$ respectively. Our results sculpt a new connection between signal propagation in neural networks and interpretability, with noise stability emerging as a powerful tool for understanding and improving modern Transformers.
Authors: Ajay Kumar Shrestha
Abstract: In federated learning (FL), decentralized model training allows multi-ple participants to collaboratively improve a shared machine learning model without exchanging raw data. However, ensuring the integrity and reliability of the system is challenging due to the presence of potentially malicious or faulty nodes that can degrade the model's performance. This paper proposes a novel trust-based incentive mechanism designed to evaluate and reward the quality of contributions in FL systems. By dynamically assessing trust scores based on fac-tors such as data quality, model accuracy, consistency, and contribution fre-quency, the system encourages honest participation and penalizes unreliable or malicious behavior. These trust scores form the basis of an incentive mechanism that rewards high-trust nodes with greater participation opportunities and penal-ties for low-trust participants. We further explore the integration of blockchain technology and smart contracts to automate the trust evaluation and incentive distribution processes, ensuring transparency and decentralization. Our proposed theoretical framework aims to create a more robust, fair, and transparent FL eco-system, reducing the risks posed by untrustworthy participants.
Authors: Madalina Erascu, Johannes Middeke
Abstract: Symmetry in integer programming causes redundant search and is often handled with symmetry breaking constraints that remove as many equivalent solutions as possible. We propose an algebraic method which allows to generate a random family of polynomial inequalities which can be used as symmetry breakers. The method requires as input an arbitrary base polynomial and a group of permutations which is specific to the integer program. The computations can be easily carried out in any major symbolic computation software. In order to test our approach, we describe a case study on near half-capacity 0-1 bin packing instances which exhibit substantial symmetries. We statically generate random quadratic breakers and add them to a baseline integer programming problem which we then solve with Gurobi. It turns out that simple symmetry breakers, especially combining few variables and permutations, most consistently reduce work time.
Authors: Nataraj Das, Atreya Vedantam, Chandrashekar Lakshminarayanan
Abstract: Grokking, the phenomenon of delayed generalization, is often attributed to the depth and compositional structure of deep neural networks. We study grokking in one of the simplest possible settings: the learning of a linear model with logistic loss for binary classification on data that are linearly (and max margin) separable about the origin. We investigate three testing regimes: (1) test data drawn from the same distribution as the training data, in which case grokking is not observed; (2) test data concentrated around the margin, in which case grokking is observed; and (3) adversarial test data generated via projected gradient descent (PGD) attacks, in which case grokking is also observed. We theoretically show that the implicit bias of gradient descent induces a three-phase learning process-population-dominated, support-vector-dominated unlearning, and support-vector-dominated generalization-during which delayed generalization can arise. Our analysis further relates the emergence of grokking to asymmetries in the data, both in the number of examples per class and in the distribution of support vectors across classes, and yields a characterization of the grokking time. We experimentally validate our theory by planting different distributions of population points and support vectors, and by analyzing accuracy curves and hyperplane dynamics. Overall, our results demonstrate that grokking does not require depth or representation learning, and can emerge even in linear models through the dynamics of the bias term.
Authors: Jared Zhu, Minhao Hu, Junde Wu
Abstract: Large language models are increasingly used as programming agents for repository level software engineering tasks. While recent benchmarks evaluate correctness in realistic codebases, they largely treat tasks as independent and do not assess whether agents can reuse experience across related problems. As a result, the ability of agents to accumulate, retrieve, and apply prior experience, as well as the efficiency gains from such reuse, remains difficult to measure. We introduce SWE-ContextBench, a benchmark designed to explicitly evaluate experience reuse in programming agents. Built on SWE-Bench Lite, SWE-ContextBench augments 300 base tasks with 99 related tasks derived from real dependency and reference relationships among GitHub issues and pull requests, forming task sequences with shared context. The benchmark evaluates agents along three complementary dimensions: prediction accuracy, time efficiency, and cost efficiency. Using SWE-ContextBench, we study multiple experience reuse settings, including oracle guided and autonomous retrieval, as well as full execution trajectories and compact summaries. Our results show that correctly selected summarized experience improves resolution accuracy and substantially reduces runtime and token cost, particularly on harder tasks. In contrast, unfiltered or incorrectly selected experience provides limited or negative benefits. These findings highlight the importance of experience representation and retrieval quality, and position SWE-ContextBench as a principled benchmark for studying experience reuse in programming agents.
Authors: Yifei Gao, Lei Wang, Rong-Cheng Tu, Qixin Zhang, Jun Cheng, Dacheng Tao
Abstract: A core bottleneck in large language model (LLM) inference is the cost of attending over the ever-growing key-value (KV) cache. Although near-oracle top-k KV selection can preserve the quality of dense attention while sharply reducing computation and bandwidth, existing sparse methods generally rely on posterior heuristics, i.e., selectors conditioned on observed attention or proxy scores. Such conditioning introduces posterior bias: it tends to distort true token importance and miss salient tokens, thereby impairing long-range reasoning. To tackle this problem, we propose Pre-hoc Sparsity (PrHS), which selects KV entries before attention scoring and provides explicit accuracy control. Let the attention mass of discarded entries be delta (the dropped mass). Through a marginal-to-mutual-information analysis, we derive an upper bound on the mutual-information loss that depends only on the dropped mass. This relation explains failure modes of posterior heuristics and enables verifiable guarantees by controlling the dropped mass in advance. Within PrHS, we instantiate three orthogonal pre-hoc selectors along the axes of time, depth, and layer. Extensive experiments on LLaMA and Mistral families validate PrHS. Across GSM8K and CoQA, PrHS reduces retrieval overhead by over 90%, achieving 3x higher retrieval sparsity than HShare at matched or better accuracy. It incurs under 1% average degradation on LongBench, lowers attention FLOPs by about 15% versus prior sparse baselines, and yields a 9.9x speedup in attention-operator latency and 2.8x higher throughput on NVIDIA A100-80GB GPUs than the dense baseline.
Authors: Ido Amos, Avi Caciularu, Mor Geva, Amir Globerson, Jonathan Herzig, Lior Shani, Idan Szpektor
Abstract: Reasoning with a chain-of-thought (CoT) enables Large Language Models (LLMs) to solve complex tasks but incurs significant inference costs due to the generation of long rationales. We propose Thinking States, a method that performs reasoning {\em while} the input is processing. Specifically, Thinking States generates sequences of thinking tokens every few input tokens, transforms the thoughts back into embedding space, and adds them to the following input tokens. This has two key advantages. First, it captures the recurrent nature of CoT, but where the thought tokens are generated as input is processing. Second, since the thoughts are represented as tokens, they can be learned from natural language supervision, and using teacher-forcing, which is parallelizable. Empirically, Thinking States outperforms other latent reasoning methods on multiple reasoning tasks, narrowing the gap to CoT on math problems, and matching its performance on 2-Hop QA with improved latency. On state-tracking tasks, we show Thinking States leads to stronger reasoning behavior than CoT, successfully extrapolating to longer sequences than seen during training.
Authors: Cristian P\'erez-Corral, Alberto Fern\'andez-Hern\'andez, Jose I. Mestre, Manuel F. Dolz, Jose Duato, Enrique S. Quintana-Ort\'i
Abstract: Despite the empirical success of DNN, their internal training dynamics remain difficult to characterize. In ReLU-based models, the activation pattern induced by a given input determines the piecewise-linear region in which the network behaves affinely. Motivated by this geometry, we investigate whether training exhibits a two-timescale behavior: an early stage with substantial changes in activation patterns and a later stage where weight updates predominantly refine the model within largely stable activation regimes. We first prove a local stability property: outside measure-zero sets of parameters and inputs, sufficiently small parameter perturbations preserve the activation pattern of a fixed input, implying locally affine behavior within activation regions. We then empirically track per-iteration changes in weights and activation patterns across fully-connected and convolutional architectures, as well as Transformer-based models, where activation patterns are recorded in the ReLU feed-forward (MLP/FFN) submodules, using fixed validation subsets. Across the evaluated settings, activation-pattern changes decay 3 times earlier than weight-update magnitudes, showing that late-stage training often proceeds within relatively stable activation regimes. These findings provide a concrete, architecture-agnostic instrument for monitoring training dynamics and motivate further study of decoupled optimization strategies for piecewise-linear networks. For reproducibility, code and experiment configurations will be released upon acceptance.
Authors: Jie Zhang, Xingtong Yu, Yuan Fang, Rudi Stouffs, Zdravko Trivic
Abstract: Learning transferable multimodal embeddings for urban environments is challenging because urban understanding is inherently spatial, yet existing datasets and benchmarks lack explicit alignment between street-view images and urban structure. We introduce UGData, a spatially grounded dataset that anchors street-view images to structured spatial graphs and provides graph-aligned supervision via spatial reasoning paths and spatial context captions, exposing distance, directionality, connectivity, and neighborhood context beyond image content. Building on UGData, we propose UGE, a two-stage training strategy that progressively and stably aligns images, text, and spatial structures by combining instruction-guided contrastive learning with graph-based spatial encoding. We finally introduce UGBench, a comprehensive benchmark to evaluate how spatially grounded embeddings support diverse urban understanding tasks -- including geolocation ranking, image retrieval, urban perception, and spatial grounding. We develop UGE on multiple state-of-the-art VLM backbones, including Qwen2-VL, Qwen2.5-VL, Phi-3-Vision, and LLaVA1.6-Mistral, and train fixed-dimensional spatial embeddings with LoRA tuning. UGE built upon Qwen2.5-VL-7B backbone achieves up to 44% improvement in image retrieval and 30% in geolocation ranking on training cities, and over 30% and 22% gains respectively on held-out cities, demonstrating the effectiveness of explicit spatial grounding for spatially intensive urban tasks.
Authors: Debajyoti Datta, Trishala Neeraj, Bibek Paudel, Vyom Sharma, Subhabrata Mukherjee
Abstract: Long-context inference is constrained by KV-cache memory, which grows linearly with sequence length; KV-cache compression therefore hinges on reliably selecting which past tokens to retain. Most geometry-based eviction methods score keys by cosine similarity to a global centroid, but cosine is scale-invariant and can discard magnitude cues that distinguish semantically salient tokens. We propose ManifoldKV, a training-free scorer that ranks tokens by Euclidean distance to the key centroid, capturing both angular and radial deviations. On the RULER benchmark, ManifoldKV achieves 95.7% accuracy at 4K-16K contexts with 20% compression; matching the best geometric baseline while improving robustness in two regimes where cosine scoring fails. First, on multi-key retrieval, ManifoldKV reduces directional collisions, achieving 92.4% vs KeyDiff's 77.0% (+15.4 points) on 3-key NIAH at 50% compression. Second, to address dilution and performance collapse of global centroids at 64K context, we introduce WindowedManifoldKV, which restores accuracy to 84.3% at 25% compression, a 49-point recovery over global L2 and +3.2 points over KeyDiff. The method requires only 3 lines of code and works across 4 architectures without tuning.
Authors: Zhiliang Chen, Alfred Wei Lun Leong, Shao Yong Ong, Apivich Hemachandram, Gregory Kang Ruey Lau, Chuan-Sheng Foo, Zhengyuan Liu, Nancy F. Chen, Bryan Kian Hsiang Low
Abstract: Co-optimizing data and model configurations for training LLMs presents a classic chicken-and-egg dilemma: The best training data configuration (e.g., data mixture) for a downstream task depends on the chosen model configuration (e.g., model architecture), and vice versa. However, jointly optimizing both data and model configurations is often deemed intractable, and existing methods focus on either data or model optimization without considering their interaction. We introduce JoBS, an approach that uses a scaling-law-inspired performance predictor to aid Bayesian optimization (BO) in jointly optimizing LLM training data and model configurations efficiently. JoBS allocates a portion of the optimization budget to learn an LLM performance predictor that predicts how promising a training configuration is from a small number of training steps. The remaining budget is used to perform BO entirely with the predictor, effectively amortizing the cost of running full-training runs. We study JoBS's average regret and devise the optimal budget allocation to minimize regret. JoBS outperforms existing multi-fidelity BO baselines, as well as data and model optimization approaches across diverse LLM tasks under the same optimization budget.
Authors: Ivan C. H. Liu, Hsiao-Yuan Chen
Abstract: Quantum mechanics occupies a central position in contemporary science while remaining largely inaccessible to direct sensory experience. This paper proposes a roadmap to quantum aesthetics that examines how quantum concepts become aesthetic phenomena through artistic mediation rather than direct representation. Two complementary and orthogonal approaches are articulated. The first, a pioneering top-down approach, employs text-prompt-based generative AI to probe quantum aesthetics as a collective cultural construct embedded in large-scale training data. By systematically modulating the linguistic weight of the term "quantum," generative models are used as experimental environments to reveal how quantum imaginaries circulate within contemporary visual culture. The second, a bottom-up approach, derives aesthetic form directly from quantum-mechanical structures through the visualization of quantum-generated data, exemplified here by hydrogen atomic orbitals calculated from the Schr\"odinger equation. These approaches are framed not as competing methods but as intersecting paths within a navigable field of artistic research. They position quantum aesthetics as an emergent field of artistic research shaped by cultural imagination, computational mediation, and physical law, opening new directions for artistic practice and pedagogy at the intersection of art, data, artificial intelligence and quantum science.
Authors: Yeke Chen, Shihao Dong, Xiaoyu Ji, Jingkai Sun, Zeren Luo, Liu Zhao, Jiahui Zhang, Wanyue Li, Ji Ma, Bowen Xu, Yimin Han, Yudong Zhao, Peng Lu
Abstract: Realizing versatile and human-like performance in high-demand sports like badminton remains a formidable challenge for humanoid robotics. Unlike standard locomotion or static manipulation, this task demands a seamless integration of explosive whole-body coordination and precise, timing-critical interception. While recent advances have achieved lifelike motion mimicry, bridging the gap between kinematic imitation and functional, physics-aware striking without compromising stylistic naturalness is non-trivial. To address this, we propose Imitation-to-Interaction, a progressive reinforcement learning framework designed to evolve a robot from a "mimic" to a capable "striker." Our approach establishes a robust motor prior from human data, distills it into a compact, model-based state representation, and stabilizes dynamics via adversarial priors. Crucially, to overcome the sparsity of expert demonstrations, we introduce a manifold expansion strategy that generalizes discrete strike points into a dense interaction volume. We validate our framework through the mastery of diverse skills, including lifts and drop shots, in simulation. Furthermore, we demonstrate the first zero-shot sim-to-real transfer of anthropomorphic badminton skills to a humanoid robot, successfully replicating the kinetic elegance and functional precision of human athletes in the physical world.
Authors: Bilgehan Sel, Vaishakh Keshava, Phillip Wallis, Lukas Rutishauser, Ming Jin, Dingcheng Li
Abstract: Addressing the critical need for robust safety in Large Language Models (LLMs), particularly against adversarial attacks and in-distribution errors, we introduce Reinforcement Learning with Backtracking Feedback (RLBF). This framework advances upon prior methods, such as BSAFE, by primarily leveraging a Reinforcement Learning (RL) stage where models learn to dynamically correct their own generation errors. Through RL with critic feedback on the model's live outputs, LLMs are trained to identify and recover from their actual, emergent safety violations by emitting an efficient "backtrack by x tokens" signal, then continuing generation autoregressively. This RL process is crucial for instilling resilience against sophisticated adversarial strategies, including middle filling, Greedy Coordinate Gradient (GCG) attacks, and decoding parameter manipulations. To further support the acquisition of this backtracking capability, we also propose an enhanced Supervised Fine-Tuning (SFT) data generation strategy (BSAFE+). This method improves upon previous data creation techniques by injecting violations into coherent, originally safe text, providing more effective initial training for the backtracking mechanism. Comprehensive empirical evaluations demonstrate that RLBF significantly reduces attack success rates across diverse benchmarks and model scales, achieving superior safety outcomes while critically preserving foundational model utility.
Authors: Zhuoen Chen, Dongfang Li, Meishan Zhang, Baotian Hu, Min Zhang
Abstract: Large Language Models (LLMs) face significant challenges in long-context processing, including quadratic computational costs, information forgetting, and the context fragmentation inherent in retrieval-augmented generation (RAG). We propose a cognitively inspired framework for efficient long-context inference based on chunk-wise compression and selective memory recall, rather than processing all raw tokens. The framework segments long inputs into chunks and encodes each chunk into compressed memory representations using a learned compressor. A gating module dynamically selects relevant memory blocks, which are then iteratively processed by a reasoning module with an evolving working memory to solve downstream tasks. The compressor and reasoner are jointly optimized via end-to-end reinforcement learning, while the gating module is trained separately as a classifier. Experimental results show that the proposed method achieves competitive accuracy on multi-hop reasoning benchmarks such as RULER-HQA, extrapolates context length from 7K to 1.75M tokens, and offers a favorable accuracy-efficiency trade-off compared to strong long-context baselines. In particular, it achieves up to a 2 times reduction in peak GPU memory usage and a 6 times inference speedup over MemAgent.
Authors: Yao-hua Franck Xu, Tayeb Lemlouma, Arnaud Braud, Jean-Marie Bonnin
Abstract: Cooperation is fundamental for society's viability, as it enables the emergence of structure within heterogeneous groups that seek collective well-being. However, individuals are inclined to defect in order to benefit from the group's cooperation without contributing the associated costs, thus leading to unfair situations. In game theory, social dilemmas entail this dichotomy between individual interest and collective outcome. The most dominant approach to multi-agent cooperation is the utilitarian welfare which can produce efficient highly inequitable outcomes. This paper proposes a novel framework to foster fairer cooperation by replacing the standard utilitarian objective with Proportional Fairness. We introduce a fair altruistic utility for each agent, defined on the individual log-payoff space and derive the analytical conditions required to ensure cooperation in classic social dilemmas. We then extend this framework to sequential settings by defining a Fair Markov Game and deriving novel fair Actor-Critic algorithms to learn fair policies. Finally, we evaluate our method in various social dilemma environments.
Authors: Xin Wu, Zhixuan Liang, Yue Ma, Mengkang Hu, Zhiyuan Qin, Xiu Li
Abstract: Multimodal Large Language Models (MLLMs) have significantly advanced embodied AI, and using them to benchmark robotic intelligence has become a pivotal trend. However, existing frameworks remain predominantly confined to single-arm manipulation, failing to capture the spatio-temporal coordination required for bimanual tasks like lifting a heavy pot. To address this, we introduce BiManiBench, a hierarchical benchmark evaluating MLLMs across three tiers: fundamental spatial reasoning, high-level action planning, and low-level end-effector control. Our framework isolates unique bimanual challenges, such as arm reachability and kinematic constraints, thereby distinguishing perceptual hallucinations from planning failures. Analysis of over 30 state-of-the-art models reveals that despite high-level reasoning proficiency, MLLMs struggle with dual-arm spatial grounding and control, frequently resulting in mutual interference and sequencing errors. These findings suggest the current paradigm lacks a deep understanding of mutual kinematic constraints, highlighting the need for future research to focus on inter-arm collision-avoidance and fine-grained temporal sequencing.
Authors: Thorsten Kl\"o{\ss}ner, Jo\~ao Belo, Zekun Wu, J\"org Hoffmann, Anna Maria Feit
Abstract: Interfaces for human oversight must effectively support users' situation awareness under time-critical conditions. We explore reinforcement learning (RL)-based UI adaptation to personalize alerting strategies that balance the benefits of highlighting critical events against the cognitive costs of interruptions. To enable learning without real-world deployment, we integrate models of users' gaze behavior to simulate attentional dynamics during monitoring. Using a delivery-drone oversight scenario, we present initial results suggesting that RL-based highlighting can outperform static, rule-based approaches and discuss challenges of intelligent oversight support.
Authors: Shujaat Khan, Waleed Iqbal Waseer
Abstract: The prediction of electromagnetic spectra for MXene-based solar absorbers is a computationally intensive task, traditionally addressed using full-wave solvers. This study introduces an efficient deep learning framework incorporating transfer learning, multi-channel spectral refinement (MCSR), and Savitzky-Golay smoothing to accelerate and enhance spectral prediction accuracy. The proposed architecture leverages a pretrained MobileNetV2 model, fine-tuned to predict 102-point absorption spectra from $64\times64$ metasurface designs. Additionally, the MCSR module processes the feature map through multi-channel convolutions, enhancing feature extraction, while Savitzky-Golay smoothing mitigates high-frequency noise. Experimental evaluations demonstrate that the proposed model significantly outperforms baseline Convolutional Neural Network (CNN) and deformable CNN models, achieving an average root mean squared error (RMSE) of 0.0245, coefficient of determination \( R^2 \) of 0.9578, and peak signal-to-noise ratio (PSNR) of 32.98 dB. The proposed framework presents a scalable and computationally efficient alternative to conventional solvers, positioning it as a viable candidate for rapid spectral prediction in nanophotonic design workflows.
Authors: Benjamin Livshits
Abstract: We argue that when it comes to producing secure code with AI, the prevailing "fighting fire with fire" approach -- using probabilistic AI-based checkers or attackers to secure probabilistically generated code -- fails to address the long tail of security bugs. As a result, systems may remain exposed to zero-day vulnerabilities that can be discovered by better-resourced or more persistent adversaries. While neurosymbolic approaches that combine LLMs with formal methods are attractive in principle, we argue that they are difficult to reconcile with the "vibe coding" workflow common in LLM-assisted development: unless the end-to-end verification pipeline is fully automated, developers are repeatedly asked to validate specifications, resolve ambiguities, and adjudicate failures, making the human-in-the-loop a likely point of weakness, compromising secure-by-construction guarantees. In this paper we argue that stronger security guarantees can be obtained by enforcing security constraints during code generation (e.g., via constrained decoding), rather than relying solely on post-hoc detection and repair. This direction is particularly promising for diffusion-style code models, whose approach provides a natural elegant opportunity for modular, hierarchical security enforcement, allowing us to combine lower-latency generation techniques with generating secure-by-construction code.
Authors: Xinghao Wang, Pengyu Wang, Xiaoran Liu, Fangxu Liu, Jason Chu, Kai Song, Xipeng Qiu
Abstract: Block-sparse attention is promising for accelerating long-context LLM pre-filling, yet identifying relevant blocks efficiently remains a bottleneck. Existing methods typically employ coarse-grained attention as a proxy for block importance estimation, but often resort to expensive token-level searching or scoring, resulting in significant selection overhead. In this work, we trace the inaccuracy of standard coarse-grained attention via mean pooling to a theoretical root cause: the interaction between mean pooling and Rotary Positional Embeddings (RoPE). We prove that mean pooling acts as a low-pass filter that induces destructive interference in high-frequency dimensions, effectively creating a "blind spot" for local positional information (e.g., slash patterns). To address this, we introduce Prism, a training-free spectral-aware approach that decomposes block selection into high-frequency and low-frequency branches. By applying energy-based temperature calibration, Prism restores the attenuated positional signals directly from pooled representations, enabling block importance estimation using purely block-level operations, thereby improving efficiency. Extensive evaluations confirm that Prism maintains accuracy parity with full attention while delivering up to $\mathbf{5.1\times}$ speedup.
Authors: Haocheng Lu, Nan Zhang, Wei Tao, Xiaoyang Qu, Guokuan Li, Jiguang Wan, Jianzong Wang
Abstract: Streaming video question answering (Streaming Video QA) poses distinct challenges for multimodal large language models (MLLMs), as video frames arrive sequentially and user queries can be issued at arbitrary time points. Existing solutions relying on fixed-size memory or naive compression often suffer from context loss or memory overflow, limiting their effectiveness in long-form, real-time scenarios. We present Vista, a novel framework for scene-aware streaming video QA that enables efficient and scalable reasoning over continuous video streams. The innovation of Vista can be summarized in three aspects: (1) scene-aware segmentation, where Vista dynamically clusters incoming frames into temporally and visually coherent scene units; (2) scene-aware compression, where each scene is compressed into a compact token representation and stored in GPU memory for efficient index-based retrieval, while full-resolution frames are offloaded to CPU memory; and (3) scene-aware recall, where relevant scenes are selectively recalled and reintegrated into the model input upon receiving a query, enabling both efficiency and completeness. Vista is model-agnostic and integrates seamlessly with a variety of vision-language backbones, enabling long-context reasoning without compromising latency or memory efficiency. Extensive experiments on StreamingBench demonstrate that Vista achieves state-of-the-art performance, establishing a strong baseline for real-world streaming video understanding.
Authors: Francesc Wilhelmi, Boris Bellalta, Miguel Casasnovas, Aleksandra Kijanka, Miguel Calvo-Fullana
Abstract: Spatial Reuse (SR) is a cost-effective technique for improving spectral efficiency in dense IEEE 802.11 deployments by enabling simultaneous transmissions. However, the decentralized optimization of SR parameters -- transmission power and Carrier Sensing Threshold (CST) -- across different Basic Service Sets (BSSs) is challenging due to the lack of global state information. In addition, the concurrent operation of multiple agents creates a highly non-stationary environment, often resulting in suboptimal global configurations (e.g., using the maximum possible transmission power by default). To overcome these limitations, this paper introduces a decentralized learning algorithm based on regret-matching, grounded in internal regret minimization. Unlike standard decentralized ``selfish'' approaches that often converge to inefficient Nash Equilibria (NE), internal regret minimization guides competing agents toward Correlated Equilibria (CE), effectively mimicking coordination without explicit communication. Through simulation results, we showcase the superiority of our proposed approach and its ability to reach near-optimal global performance. These results confirm the not-yet-unleashed potential of scalable decentralized solutions and question the need for the heavy signaling overheads and architectural complexity associated with emerging centralized solutions like Multi-Access Point Coordination (MAPC).
Authors: Alif Rizqullah Mahdi, Mahdi Rezaei, Natasha Merat
Abstract: Gestures are a key component of non-verbal communication in traffic, often helping pedestrian-to-driver interactions when formal traffic rules may be insufficient. This problem becomes more apparent when autonomous vehicles (AVs) struggle to interpret such gestures. In this study, we present a gesture classification framework using 2D pose estimation applied to real-world video sequences from the WIVW dataset. We categorise gestures into four primary classes (Stop, Go, Thank & Greet, and No Gesture) and extract 76 static and dynamic features from normalised keypoints. Our analysis demonstrates that hand position and movement velocity are especially discriminative in distinguishing between gesture classes, achieving a classification accuracy score of 87%. These findings not only improve the perceptual capabilities of AV systems but also contribute to the broader understanding of pedestrian behaviour in traffic contexts.
Authors: Hengyu Liu, Tianyi Li, Haoyu Wang, Kristian Torp, Yushuai Li, Tiancheng Zhang, Torben Bach Pedersen, Christian S. Jensen
Abstract: Vessel trajectory data from the Automatic Identification System (AIS) is used widely in maritime analytics. Yet, analysis is difficult for non-expert users due to the incompleteness and complexity of AIS data. We present CLEAR, a knowledge-centric vessel trajectory analysis platform that aims to overcome these barriers. By leveraging the reasoning and generative capabilities of Large Language Models (LLMs), CLEAR transforms raw AIS data into complete, interpretable, and easily explorable vessel trajectories through a Structured Data-derived Knowledge Graph (SD-KG). As part of the demo, participants can configure parameters to automatically download and process AIS data, observe how trajectories are completed and annotated, inspect both raw and imputed segments together with their SD-KG evidence, and interactively explore the SD-KG through a dedicated graph viewer, gaining an intuitive and transparent understanding of vessel movements.
Authors: Xiaodong Lu, Xiaohan Wang, Jiajun Chai, Guojun Yin, Wei Lin, Zhijun Chen, Yu Luo, Fuzhen Zhuang, Yikun Ban, Deqing Wang
Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) is an effective paradigm for improving the reasoning capabilities of large language models. However, existing RLVR methods utilize rollouts in an indiscriminate and short-horizon manner: responses of heterogeneous quality within each prompt are treated uniformly, and historical rollouts are discarded after a single use. This leads to noisy supervision, poor sample efficiency, and suboptimal policy updates. We address these issues by formulating rollout scheduling in RLVR as a contextual bandit problem and proposing a unified neural scheduling framework that adaptively selects high-value rollouts throughout training. Each rollout is treated as an arm whose reward is defined by the induced performance gain between consecutive optimization steps. The resulting scheduler supports both noise-aware intra-group selection and adaptive global reuse of historical rollouts within a single principled framework. We provide theoretical justification by deriving sublinear regret bounds and showing that enlarging the rollout buffer improves the achievable performance upper bound. Experiments on six mathematical reasoning benchmarks demonstrate consistent gains in performance and training efficiency across multiple RLVR optimization methods.
Authors: Piotr Skowron
Abstract: We consider a model where a subset of candidates must be selected based on voter preferences, subject to general constraints that specify which subsets are feasible. This model generalizes committee elections with diversity constraints, participatory budgeting (including constraints specifying how funds must be allocated to projects from different pools), and public decision-making. Axioms of proportionality have recently been defined for this general model, but the proposed rules apply only to approval ballots, where each voter submits a subset of candidates she finds acceptable. We propose proportional rules for cardinal ballots, where each voter assigns a numerical value to each candidate corresponding to her utility if that candidate is selected. In developing these rules, we also introduce methods that produce proportional rankings, ensuring that every prefix of the ranking satisfies proportionality.
Authors: Yutao Zhu, Xingshuo Zhang, Maosen Zhang, Jiajie Jin, Liancheng Zhang, Xiaoshuai Song, Kangzhi Zhao, Wencong Zeng, Ruiming Tang, Han Li, Ji-Rong Wen, Zhicheng Dou
Abstract: The advancement of large language models (LLMs) has significantly accelerated the development of search agents capable of autonomously gathering information through multi-turn web interactions. Various benchmarks have been proposed to evaluate such agents. However, existing benchmarks often construct queries backward from answers, producing unnatural tasks misaligned with real-world needs. Moreover, these benchmarks tend to focus on either locating specific information or aggregating information from multiple sources, while relying on static answer sets prone to data contamination. To bridge these gaps, we introduce GISA, a benchmark for General Information-Seeking Assistants comprising 373 human-crafted queries that reflect authentic information-seeking scenarios. GISA features four structured answer formats (item, set, list, and table), enabling deterministic evaluation. It integrates both deep reasoning and broad information aggregation within unified tasks, and includes a live subset with periodically updated answers to resist memorization. Notably, GISA provides complete human search trajectories for every query, offering gold-standard references for process-level supervision and imitation learning. Experiments on mainstream LLMs and commercial search products reveal that even the best-performing model achieves only 19.30\% exact match score, with performance notably degrading on tasks requiring complex planning and comprehensive information gathering. These findings highlight substantial room for future improvement.
Authors: Shih-Fang Chen, Jun-Cheng Chen, I-Hong Jhuo, Yen-Yu Lin
Abstract: Human perception for effective object tracking in a 2D video stream arises from the implicit use of prior 3D knowledge combined with semantic reasoning. In contrast, most generic object tracking (GOT) methods primarily rely on 2D features of the target and its surroundings while neglecting 3D geometric cues, which makes them susceptible to partial occlusion, distractors, and variations in geometry and appearance. To address this limitation, we introduce GOT-Edit, an online cross-modality model editing approach that integrates geometry-aware cues into a generic object tracker from a 2D video stream. Our approach leverages features from a pre-trained Visual Geometry Grounded Transformer to enable geometric cue inference from only a few 2D images. To tackle the challenge of seamlessly combining geometry and semantics, GOT-Edit performs online model editing with null-space constrained updates that incorporate geometric information while preserving semantic discrimination, yielding consistently better performance across diverse scenarios. Extensive experiments on multiple GOT benchmarks demonstrate that GOT-Edit achieves superior robustness and accuracy, particularly under occlusion and clutter, establishing a new paradigm for combining 2D semantics with 3D geometric reasoning for generic object tracking.
Authors: Ahmed Salem, Andrew Paverd, Sahar Abdelnabi
Abstract: Large language models (LLMs) are commonly treated as stateless: once an interaction ends, no information is assumed to persist unless it is explicitly stored and re-supplied. We challenge this assumption by introducing implicit memory-the ability of a model to carry state across otherwise independent interactions by encoding information in its own outputs and later recovering it when those outputs are reintroduced as input. This mechanism does not require any explicit memory module, yet it creates a persistent information channel across inference requests. As a concrete demonstration, we introduce a new class of temporal backdoors, which we call time bombs. Unlike conventional backdoors that activate on a single trigger input, time bombs activate only after a sequence of interactions satisfies hidden conditions accumulated via implicit memory. We show that such behavior can be induced today through straightforward prompting or fine-tuning. Beyond this case study, we analyze broader implications of implicit memory, including covert inter-agent communication, benchmark contamination, targeted manipulation, and training-data poisoning. Finally, we discuss detection challenges and outline directions for stress-testing and evaluation, with the goal of anticipating and controlling future developments. To promote future research, we release code and data at: https://github.com/microsoft/implicitMemory.
Authors: Leon Fr\"ohling, Alessandro Giaconia, Edyta Paulina Bogucka, Daniele Quercia
Abstract: AI impact assessments often stress near-term risks because human judgment degrades over longer horizons, exemplifying the Collingridge dilemma: foresight is most needed when knowledge is scarcest. To address long-term systemic risks, we introduce a scalable approach that simulates in-silico agents using the strategic foresight method of the Futures Wheel. We applied it to four AI uses spanning Technology Readiness Levels (TRLs): Chatbot Companion (TRL 9, mature), AI Toy (TRL 7, medium), Griefbot (TRL 5, low), and Death App (TRL 2, conceptual). Across 30 agent runs per use, agents produced 86-110 consequences, condensed into 27-47 unique risks. To benchmark the agent outputs against human perspectives, we collected evaluations from 290 domain experts and 7 leaders, and conducted Futures Wheel sessions with 42 experts and 42 laypeople. Agents generated many systemic consequences across runs. Compared with these outputs, experts identified fewer risks, typically less systemic but judged more likely, whereas laypeople surfaced more emotionally salient concerns that were generally less systemic. We propose a hybrid foresight workflow, wherein agents broaden systemic coverage, and humans provide contextual grounding. Our dataset is available at: https://social-dynamics.net/ai-risks/foresight.
Authors: Ziyao Tang, Pengkun Jiao, Xinhang Chen, Wei Liu, Shiyong Li, Jingjing Chen
Abstract: Given the quadratic complexity of attention, KV cache eviction is vital to accelerate model inference. Current KV cache eviction methods typically rely on instantaneous heuristic metrics, implicitly assuming that score magnitudes are consistent proxies for importance across all heads. However, this overlooks the heterogeneity in predictive fidelity across attention heads. While certain heads prioritize the instantaneous contribution of tokens, others are dedicated to capturing long-horizon utility. In this paper, we propose that optimal budget allocation should be governed by the marginal utility in preserving long-term semantic information. Based on this insight, we propose LU-KV, a novel framework that optimizes head-level budget allocation through a convex-hull relaxation and a marginal-utility-based greedy solver to achieve near-optimal precision. Furthermore, we implement a data-driven offline profiling protocol to facilitate the practical deployment of LU-KV. Extensive evaluations on LongBench and RULER benchmarks demonstrate that LU-KV achieves an 80% reduction in KV cache size with minimal performance degradation, while simultaneously reducing inference latency and GPU memory footprint.
Authors: Muhammad Saad Ali, Daanish U. Khan, Laiba Intizar Ahmad, Umer Irfan, Maryam Mustafa, Naveed Anwar Bhatti, Muhammad Hamad Alizai
Abstract: We present Kissan-Dost, a multilingual, sensor-grounded conversational system that turns live on-farm measurements and weather into plain-language guidance delivered over WhatsApp text or voice. The system couples commodity soil and climate sensors with retrieval-augmented generation, then enforces grounding, traceability, and proactive alerts through a modular pipeline. In a 90-day, two-site pilot with five participants, we ran three phases (baseline, dashboard only, chatbot only). Dashboard engagement was sporadic and faded, while the chatbot was used nearly daily and informed concrete actions. Controlled tests on 99 sensor-grounded crop queries achieved over 90 percent correctness with subsecond end-to-end latency, alongside high-quality translation outputs. Results show that careful last-mile integration, not novel circuitry, unlocks the latent value of existing Agri-IoT for smallholders.
Authors: Heiko Hoppe, Fabian Akkerman, Wouter van Heeswijk, Maximilian Schiffer
Abstract: Reinforcement Learning is increasingly applied to logistics, scheduling, and recommender systems, but standard algorithms struggle with the curse of dimensionality in such large discrete action spaces. Existing algorithms typically rely on restrictive grid-based structures or computationally expensive nearest-neighbor searches, limiting their effectiveness in high-dimensional or irregularly structured domains. We propose Distance-Guided Reinforcement Learning (DGRL), combining Sampled Dynamic Neighborhoods (SDN) and Distance-Based Updates (DBU) to enable efficient RL in spaces with up to 10$^\text{20}$ actions. Unlike prior methods, SDN leverages a semantic embedding space to perform stochastic volumetric exploration, provably providing full support over a local trust region. Complementing this, DBU transforms policy optimization into a stable regression task, decoupling gradient variance from action space cardinality and guaranteeing monotonic policy improvement. DGRL naturally generalizes to hybrid continuous-discrete action spaces without requiring hierarchical dependencies. We demonstrate performance improvements of up to 66% against state-of-the-art benchmarks across regularly and irregularly structured environments, while simultaneously improving convergence speed and computational complexity.
Authors: Laura-Maria Cornei, Mihaela-Elena Breab\u{a}n
Abstract: This paper investigates the impact of hybridizing a multi-modal Genetic Algorithm with a Graph Neural Network for timetabling optimization. The Graph Neural Network is designed to encapsulate general domain knowledge to improve schedule quality, while the Genetic Algorithm explores different regions of the search space and integrates the deep learning model as an enhancement operator to guide the solution search towards optimality. Initially, both components of the hybrid technique were designed, developed, and optimized independently to solve the tackled task. Multiple experiments were conducted on Staff Rostering, a well-known timetabling problem, to compare the proposed hybridization with the standalone optimized versions of the Genetic Algorithm and Graph Neural Network. The experimental results demonstrate that the proposed hybridization brings statistically significant improvements in both the time efficiency and solution quality metrics, compared to the standalone methods. To the best of our knowledge, this work proposes the first hybridization of a Genetic Algorithm with a Graph Neural Network for solving timetabling problems.
Authors: Yukun Jiang, Hai Huang, Mingjie Li, Yage Zhang, Michael Backes, Yang Zhang
Abstract: By introducing routers to selectively activate experts in Transformer layers, the mixture-of-experts (MoE) architecture significantly reduces computational costs in large language models (LLMs) while maintaining competitive performance, especially for models with massive parameters. However, prior work has largely focused on utility and efficiency, leaving the safety risks associated with this sparse architecture underexplored. In this work, we show that the safety of MoE LLMs is as sparse as their architecture by discovering unsafe routes: routing configurations that, once activated, convert safe outputs into harmful ones. Specifically, we first introduce the Router Safety importance score (RoSais) to quantify the safety criticality of each layer's router. Manipulation of only the high-RoSais router(s) can flip the default route into an unsafe one. For instance, on JailbreakBench, masking 5 routers in DeepSeek-V2-Lite increases attack success rate (ASR) by over 4$\times$ to 0.79, highlighting an inherent risk that router manipulation may naturally occur in MoE LLMs. We further propose a Fine-grained token-layer-wise Stochastic Optimization framework to discover more concrete Unsafe Routes (F-SOUR), which explicitly considers the sequentiality and dynamics of input tokens. Across four representative MoE LLM families, F-SOUR achieves an average ASR of 0.90 and 0.98 on JailbreakBench and AdvBench, respectively. Finally, we outline defensive perspectives, including safety-aware route disabling and router training, as promising directions to safeguard MoE LLMs. We hope our work can inform future red-teaming and safeguarding of MoE LLMs. Our code is provided in https://github.com/TrustAIRLab/UnsafeMoE.
Authors: Bo Peng, Sirui Chen, Jiaguo Tian, Yu Qiao, Chaochao Lu
Abstract: Causal discovery is essential for advancing data-driven fields such as scientific AI and data analysis, yet existing approaches face significant time- and space-efficiency bottlenecks when scaling to large graphs. To address this challenge, we present CauScale, a neural architecture designed for efficient causal discovery that scales inference to graphs with up to 1000 nodes. CauScale improves time efficiency via a reduction unit that compresses data embeddings and improves space efficiency by adopting tied attention weights to avoid maintaining axis-specific attention maps. To keep high causal discovery accuracy, CauScale adopts a two-stream design: a data stream extracts relational evidence from high-dimensional observations, while a graph stream integrates statistical graph priors and preserves key structural signals. CauScale successfully scales to 500-node graphs during training, where prior work fails due to space limitations. Across testing data with varying graph scales and causal mechanisms, CauScale achieves 99.6% mAP on in-distribution data and 84.4% on out-of-distribution data, while delivering 4-13,000 times inference speedups over prior methods. Our project page is at https://github.com/OpenCausaLab/CauScale.
Authors: Adi Haviv, Niva Elkin-Koren, Uri Hacohen, Roi Livni, Shay Moran
Abstract: The widespread use of foundation models has introduced a new risk factor of copyright issue. This issue is leading to an active, lively and on-going debate amongst the data-science community as well as amongst legal scholars. Where claims and results across both sides are often interpreted in different ways and leading to different implications. Our position is that much of the technical literature relies on traditional reconstruction techniques that are not designed for copyright analysis. As a result, memorization and copying have been conflated across both technical and legal communities and in multiple contexts. We argue that memorization, as commonly studied in data science, should not be equated with copying and should not be used as a proxy for copyright infringement. We distinguish technical signals that meaningfully indicate infringement risk from those that instead reflect lawful generalization or high-frequency content. Based on this analysis, we advocate for an output-level, risk-based evaluation process that aligns technical assessments with established copyright standards and provides a more principled foundation for research, auditing, and policy.
Authors: Dezheng Wang, Tong Chen, Guansong Pang, Congyan Chen, Shihua Li, Hongzhi Yin
Abstract: As a fundamental data mining task, unsupervised time series anomaly detection (TSAD) aims to build a model for identifying abnormal timestamps without assuming the availability of annotations. A key challenge in unsupervised TSAD is that many anomalies are too subtle to exhibit detectable deviation in any single view (e.g., time domain), and instead manifest as inconsistencies across multiple views like time, frequency, and a mixture of resolutions. However, most cross-view methods rely on feature or score fusion and do not enforce analysis-synthesis consistency, meaning the frequency branch is not required to reconstruct the time signal through an inverse transform, and vice versa. In this paper, we present Learnable Fusion of Tri-view Tokens (LEFT), a unified unsupervised TSAD framework that models anomalies as inconsistencies across complementary representations. LEFT learns feature tokens from three views of the same input time series: frequency-domain tokens that embed periodicity information, time-domain tokens that capture local dynamics, and multi-scale tokens that learns abnormal patterns at varying time series granularities. By learning a set of adaptive Nyquist-constrained spectral filters, the original time series is rescaled into multiple resolutions and then encoded, allowing these multi-scale tokens to complement the extracted frequency- and time-domain information. When generating the fused representation, we introduce a novel objective that reconstructs fine-grained targets from coarser multi-scale structure, and put forward an innovative time-frequency cycle consistency constraint to explicitly regularize cross-view agreement. Experiments on real-world benchmarks show that LEFT yields the best detection accuracy against SOTA baselines, while achieving a 5x reduction on FLOPs and 8x speed-up for training.
Authors: Alexandre Verine, Rafael Pinot, Florian Le Bronnec
Abstract: Fairness is a crucial concern for generative models, which not only reflect but can also amplify societal and cultural biases. Existing fairness notions for generative models are largely adapted from classification and focus on balancing the probability of generating samples from each sensitive group. We show that such criteria are brittle, as they can be met even when different sensitive groups are modeled with widely varying quality. To address this limitation, we introduce a new fairness definition for generative models, termed as equalized generative treatment (EGT), which requires comparable generation quality across all sensitive groups, with quality measured via a reference f-divergence. We further analyze the trade-offs induced by EGT, demonstrating that enforcing fairness constraints necessarily couples the overall model quality to that of the most challenging group to approximate. This indicates that a simple yet efficient min-max fine-tuning method should be able to balance f-divergences across sensitive groups to satisfy EGT. We validate this theoretical insight through a set of experiments on both image and text generation tasks. We demonstrate that min-max methods consistently achieve fairer outcomes compared to other approaches from the literature, while maintaining competitive overall performance for both tasks.
Authors: Mohamed Amine Ferrag, Abderrahmane Lakas, Merouane Debbah
Abstract: This paper introduces 6G-Bench, an open benchmark for evaluating semantic communication and network-level reasoning in AI-native 6G networks. 6G-Bench defines a taxonomy of 30 decision-making tasks (T1--T30) extracted from ongoing 6G and AI-agent standardization activities in 3GPP, IETF, ETSI, ITU-T, and the O-RAN Alliance, and organizes them into five standardization-aligned capability categories. Starting from 113,475 scenarios, we generate a balanced pool of 10,000 very-hard multiple-choice questions using task-conditioned prompts that enforce multi-step quantitative reasoning under uncertainty and worst-case regret minimization over multi-turn horizons. After automated filtering and expert human validation, 3,722 questions are retained as a high-confidence evaluation set, while the full pool is released to support training and fine-tuning of 6G-specialized models. Using 6G-Bench, we evaluate 22 foundation models spanning dense and mixture-of-experts architectures, short- and long-context designs (up to 1M tokens), and both open-weight and proprietary systems. Across models, deterministic single-shot accuracy (pass@1) spans a wide range from 0.22 to 0.82, highlighting substantial variation in semantic reasoning capability. Leading models achieve intent and policy reasoning accuracy in the range 0.87--0.89, while selective robustness analysis on reasoning-intensive tasks shows pass@5 values ranging from 0.20 to 0.91. To support open science and reproducibility, we release the 6G-Bench dataset on GitHub: https://github.com/maferrag/6G-Bench
Authors: Tiwei Bie, Maosong Cao, Xiang Cao, Bingsen Chen, Fuyuan Chen, Kun Chen, Lun Du, Daozhuo Feng, Haibo Feng, Mingliang Gong, Zhuocheng Gong, Yanmei Gu, Jian Guan, Kaiyuan Guan, Hongliang He, Zenan Huang, Juyong Jiang, Zhonghui Jiang, Zhenzhong Lan, Chengxi Li, Jianguo Li, Zehuan Li, Huabin Liu, Lin Liu, Guoshan Lu, Yuan Lu, Yuxin Ma, Xingyu Mou, Zhenxuan Pan, Kaida Qiu, Yuji Ren, Jianfeng Tan, Yiding Tian, Zian Wang, Lanning Wei, Tao Wu, Yipeng Xing, Wentao Ye, Liangyu Zha, Tianze Zhang, Xiaolu Zhang, Junbo Zhao, Da Zheng, Hao Zhong, Wanli Zhong, Jun Zhou, Junlin Zhou, Liwang Zhu, Muzhi Zhu, Yihong Zhuang
Abstract: While LLaDA2.0 showcased the scaling potential of 100B-level block-diffusion models and their inherent parallelization, the delicate equilibrium between decoding speed and generation quality has remained an elusive frontier. Today, we unveil LLaDA2.1, a paradigm shift designed to transcend this trade-off. By seamlessly weaving Token-to-Token (T2T) editing into the conventional Mask-to-Token (M2T) scheme, we introduce a joint, configurable threshold-decoding scheme. This structural innovation gives rise to two distinct personas: the Speedy Mode (S Mode), which audaciously lowers the M2T threshold to bypass traditional constraints while relying on T2T to refine the output; and the Quality Mode (Q Mode), which leans into conservative thresholds to secure superior benchmark performances with manageable efficiency degrade. Furthering this evolution, underpinned by an expansive context window, we implement the first large-scale Reinforcement Learning (RL) framework specifically tailored for dLLMs, anchored by specialized techniques for stable gradient estimation. This alignment not only sharpens reasoning precision but also elevates instruction-following fidelity, bridging the chasm between diffusion dynamics and complex human intent. We culminate this work by releasing LLaDA2.1-Mini (16B) and LLaDA2.1-Flash (100B). Across 33 rigorous benchmarks, LLaDA2.1 delivers strong task performance and lightning-fast decoding speed. Despite its 100B volume, on coding tasks it attains an astounding 892 TPS on HumanEval+, 801 TPS on BigCodeBench, and 663 TPS on LiveCodeBench.
Authors: Ning Yang, Chengzhi Wang, Yibo Liu, Baoliang Tian, Haijun Zhang
Abstract: Large Language Models (LLMs) in long-context scenarios are severely constrained by the linear growth of Key-Value (KV) cache memory. Existing KV compression methods rely either on static thresholds and attention-only heuristics or on coarse memory budget allocation. Under tight memory budgets, these methods overlook two key factors: prompt-dependent variation in compression risk and functional heterogeneity across attention heads, which destabilize token selection and lead to tail failures. To address these challenges, we propose CompilerKV, a risk-adaptive and head-aware compression framework that compiles offline experience into reusable decision tables for prefill-only deployment. CompilerKV integrates two key synergistic components: (i) a Head Heterogeneity Table, learned via offline contextual bandits, which assigns head-specific reliability weights to govern functional differences across attention heads explicitly; and (ii) a Risk-Adaptive Threshold Gating mechanism that jointly models attention entropy and local perplexity, transforming prompt-level risk into deployable retention thresholds. Experiments on LongBench show CompilerKV dominates SOTA methods under a 512-token budget, recovering 97.7\% of FullKV performance while achieving up to +5.2 points gain over the strongest competitor.
Authors: Stefan Szeider
Abstract: We present PBLean, a method for importing VeriPB pseudo-Boolean (PB) proof certificates into Lean 4. Key to our approach is reflection: a Boolean checker function whose soundness is fully proved in Lean and executed as compiled native code. Our method scales to proofs with tens of thousands of steps that would exhaust memory under explicit proof-term construction. Our checker supports all VeriPB kernel rules, including cutting-plane derivations and proof-by-contradiction subproofs. In contrast to external verified checkers that produce verdicts, our integration yields Lean theorems that can serve as composable lemmas in larger formal developments. To derive theorems about the original combinatorial problems rather than about PB constraints alone, we support verified encodings. This closes the trust gap between solver output and problem semantics since the constraint translation and its correctness proof are both formalized in Lean. We demonstrate the approach on various combinatorial problems.
Authors: Alexandra Pregent
Abstract: The prospect of AI systems that I call ideal emotion recognition technologies (ERTs) is often defended on the assumption that social life would benefit from increased affective transparency. This paper challenges that assumption by examining the technosocial risks posed by ideal ERTs, understood as multimodal systems capable of reliably inferring inner affective states in real time. Drawing on philosophical accounts of emotional expression and social practice, as well as empirical work in affective science and social psychology, I argue that the appeal of such systems rests on a misunderstanding of the social functions of emotional expression. Emotional expressions function not only as read-outs of inner states, but also as tools for coordinating action, enabling moral repair, sustaining interpersonal trust, and supporting collective norms. These functions depend on a background of partial opacity and epistemic friction. When deployed in socially authoritative or evaluative contexts, ideal ERTs threaten this expressive space by collapsing epistemic friction, displacing relational meaning with technology-mediated affective profiles, and narrowing the space for aspirational and role-sensitive expressions. The result is a drift towards affective determinism and ambient forms of affective auditing, which undermine both social cohesion and individual agency. I argue that, although it is intuitive to think that increasing accuracy would legitimise such systems, in the case of ERTs accuracy does not straightforwardly justify their deployment, and may, in some contexts, provide a reason for regulatory restraint. I conclude by defending a function-first regulatory approach that treats expressive discretion and intentional emotional expression as constitutive of certain social goods, and that accordingly seeks to protect these goods from excessive affective legibility.
Authors: Farnaz Khun Jush, Grit Werner, Mark Klemens, Matthias Lenga
Abstract: Reliable identification of anatomical body regions is a prerequisite for many automated medical imaging workflows, yet existing solutions remain heavily dependent on unreliable DICOM metadata. Current solutions mainly use supervised learning, which limits their applicability in many real-world scenarios. In this work, we investigate whether body region detection in volumetric CT and MR images can be achieved in a fully zero-shot manner by using knowledge embedded in large pre-trained foundation models. We propose and systematically evaluate three training-free pipelines: (1) a segmentation-driven rule-based system leveraging pre-trained multi-organ segmentation models, (2) a Multimodal Large Language Model (MLLM) guided by radiologist-defined rules, and (3) a segmentation-aware MLLM that combines visual input with explicit anatomical evidence. All methods are evaluated on 887 heterogeneous CT and MR scans with manually verified anatomical region labels. The segmentation-driven rule-based approach achieves the strongest and most consistent performance, with weighted F1-scores of 0.947 (CT) and 0.914 (MR), demonstrating robustness across modalities and atypical scan coverage. The MLLM performs competitively in visually distinctive regions, while the segmentation-aware MLLM reveals fundamental limitations.
Authors: Dalton Jones, Junyoung Park, Matthew Morse, Mingu Lee, Chris Lott, Harper Langston
Abstract: We present QUOKA: Query-oriented KV selection for efficient attention, a training-free and hardware agnostic sparse attention algorithm for accelerating transformer inference under chunked prefill. While many queries focus on a smaller group of keys in the attention operator, we observe that queries with low cosine similarity with respect to the mean query interact more strongly with more keys and have the greatest contribution to final attention logits. By prioritizing these low cosine similarity queries, the behavior of full attention during the prefill stage can be closely approximated. QUOKA leverages this observation, accelerating attention by (1) first retaining a small set of representative queries and (2) then subselectin the keys most aligned with those queries. Through experiments on Needle-In-A-Haystack, LongBench, RULER, and Math500, we show that, while realizing a 3x reduction in time-to-first-token, 5x speedup in attention on Nvidia GPUs and up to nearly a 7x speedup on Intel Xeon CPUs, QUOKA achieves near-baseline accuracy, utilizing 88% fewer key-value pairs per attention evaluation.
Authors: Johannes Thalhammer, Tina Dorosti, Sebastian Peterhansl, Daniela Pfeiffer, Franz Pfeiffer, Florian Schaff
Abstract: Undersampled CT volumes minimize acquisition time and radiation exposure but introduce artifacts degrading image quality and diagnostic utility. Reducing these artifacts is critical for high-quality imaging. We propose a computationally efficient hybrid deep-learning framework that combines the strengths of 2D and 3D models. First, a 2D U-Net operates on individual slices of undersampled CT volumes to extract feature maps. These slice-wise feature maps are then stacked across the volume and used as input to a 3D decoder, which utilizes contextual information across slices to predict an artifact-free 3D CT volume. The proposed two-stage approach balances the computational efficiency of 2D processing with the volumetric consistency provided by 3D modeling. The results show substantial improvements in inter-slice consistency in coronal and sagittal direction with low computational overhead. This hybrid framework presents a robust and efficient solution for high-quality 3D CT image post-processing. The code of this project can be found on github: https://github.com/J-3TO/2D-3DCNN_sparseview/.
Authors: Saku Peltonen, Roger Wattenhofer
Abstract: Machine learning approaches to solving Boolean Satisfiability (SAT) aim to replace handcrafted heuristics with learning-based models. Graph Neural Networks have emerged as the main architecture for SAT solving, due to the natural graph representation of Boolean formulas. We analyze the expressive power of GNNs for SAT solving through the lens of the Weisfeiler-Leman (WL) test. As our main result, we prove that the full WL hierarchy cannot, in general, distinguish between satisfiable and unsatisfiable instances. We show that indistinguishability under higher-order WL carries over to practical limitations for WL-bounded solvers that set variables sequentially. We further study the expressivity required for several important families of SAT instances, including regular, random and planar instances. To quantify expressivity needs in practice, we conduct experiments on random instances from the G4SAT benchmark and industrial instances from the International SAT Competition. Our results suggest that while random instances are largely distinguishable, industrial instances often require more expressivity to predict a satisfying assignment.
Authors: Hjalti Thrastarson, Lotta M. Ellingsen
Abstract: Skull stripping magnetic resonance images (MRI) of the human brain is an important process in many image processing techniques, such as automatic segmentation of brain structures. Numerous methods have been developed to perform this task, however, they often fail in the presence of neuropathology and can be inconsistent in defining the boundary of the brain mask. Here, we propose a novel approach to skull strip T1-weighted images in a robust and efficient manner, aiming to consistently segment the outer surface of the brain, including the sulcal cerebrospinal fluid (CSF), while excluding the full extent of the subarachnoid space and meninges. We train a modified version of the U-net on silver-standard ground truth data using a novel loss function based on the signed-distance transform (SDT). We validate our model both qualitatively and quantitatively using held-out data from the training dataset, as well as an independent external dataset. The brain masks used for evaluation partially or fully include the subarachnoid space, which may introduce bias into the comparison; nonetheless, our model demonstrates strong performance on the held-out test data, achieving a consistent mean Dice similarity coefficient (DSC) of 0.964$\pm$0.006 and an average symmetric surface distance (ASSD) of 1.4mm$\pm$0.2mm. Performance on the external dataset is comparable, with a DSC of 0.958$\pm$0.006 and an ASSD of 1.7$\pm$0.2mm. Our method achieves performance comparable to or better than existing state-of-the-art methods for brain extraction, particularly in its highly consistent preservation of the brain's outer surface. The method is publicly available on GitHub.
Authors: Micah Villmow
Abstract: LLM-based tools are automating more software development tasks at a rapid pace, but there is no rigorous way to evaluate how different architectural choices -- prompts, skills, tools, multi-agent setups -- materially affect both capability and cost. This paper introduces Scylla, an evaluation framework for benchmarking agentic coding tools through structured ablation studies that uses seven testing tiers (T0-T6) progressively adding complexity to isolate what directly influences results and how. The key metric is Cost-of-Pass (CoP): the expected dollar cost to get one correct solution, which directly quantifies the trade-off between complexity and efficiency. The framework is model-agnostic, designed to work with any CLI tool; this paper demonstrates it with Claude Sonnet 4.5, using multiple LLM judges (Opus 4.5, Sonnet 4.5, Haiku 4.5) from the same vendor for evaluation consensus, where judges score results using direct tests, human-designed LLM-evaluated rubrics, and qualitative assessment. The result is a reproducible framework that quantifies trade-offs between agent complexity and actual outcomes, suggesting that architectural complexity does not always improve quality.
Authors: Chi-Sheng Chen, Xinyu Zhang, En-Jui Kuo, Guan-Ying Chen, Qiuzhe Xie, Fan Zhang
Abstract: Time series forecasting models often lack interpretability, limiting their adoption in domains requiring explainable predictions. We propose \textsc{FreqLens}, an interpretable forecasting framework that discovers and attributes predictions to learnable frequency components. \textsc{FreqLens} introduces two key innovations: (1) \emph{learnable frequency discovery} -- frequency bases are parameterized via sigmoid mapping and learned from data with diversity regularization, enabling automatic discovery of dominant periodic patterns without domain knowledge; and (2) \emph{axiomatic frequency attribution} -- a theoretically grounded framework that provably satisfies Completeness, Faithfulness, Null-Frequency, and Symmetry axioms, with per-frequency attributions equivalent to Shapley values. On Traffic and Weather datasets, \textsc{FreqLens} achieves competitive or superior performance while discovering physically meaningful frequencies: all 5 independent runs discover the 24-hour daily cycle ($24.6 \pm 0.1$h, 2.5\% error) and 12-hour half-daily cycle ($11.8 \pm 0.1$h, 1.6\% error) on Traffic, and weekly cycles ($10\times$ longer than the input window) on Weather. These results demonstrate genuine frequency-level knowledge discovery with formal theoretical guarantees on attribution quality.
Authors: Nicol\'as Villagr\'an Prieto, Eduardo C. Garrido-Merch\'an
Abstract: Bayesian Optimization (BO) is a standard tool for hyperparameter tuning thanks to its sample efficiency on expensive black-box functions. While most BO pipelines begin with uniform random initialization, default hyperparameter values shipped with popular ML libraries such as scikit-learn encode implicit expert knowledge and could serve as informative starting points that accelerate convergence. This hypothesis, despite its intuitive appeal, has remained largely unexamined. We formalize the idea by initializing BO with points drawn from truncated Gaussian distributions centered at library defaults and compare the resulting trajectories against a uniform-random baseline. We conduct an extensive empirical evaluation spanning three BO back-ends (BoTorch, Optuna, Scikit-Optimize), three model families (Random Forests, Support Vector Machines, Multilayer Perceptrons), and five benchmark datasets covering classification and regression tasks. Performance is assessed through convergence speed and final predictive quality, and statistical significance is determined via one-sided binomial tests. Across all conditions, default-informed initialization yields no statistically significant advantage over purely random sampling, with p-values ranging from 0.141 to 0.908. A sensitivity analysis on the prior variance confirms that, while tighter concentration around the defaults improves early evaluations, this transient benefit vanishes as optimization progresses, leaving final performance unchanged. Our results provide no evidence that default hyperparameters encode useful directional information for optimization. We therefore recommend that practitioners treat hyperparameter tuning as an integral part of model development and favor principled, data-driven search strategies over heuristic reliance on library defaults.
Authors: Hao Dong, Eleni Chatzi, Olga Fink
Abstract: The pantograph-catenary interface is essential for ensuring uninterrupted and reliable power delivery in electrified rail systems. However, electrical arcing at this interface poses serious risks, including accelerated wear of contact components, degraded system performance, and potential service disruptions. Detecting arcing events at the pantograph-catenary interface is challenging due to their transient nature, noisy operating environment, data scarcity, and the difficulty of distinguishing arcs from other similar transient phenomena. To address these challenges, we propose a novel multimodal framework that combines high-resolution image data with force measurements to more accurately and robustly detect arcing events. First, we construct two arcing detection datasets comprising synchronized visual and force measurements. One dataset is built from data provided by the Swiss Federal Railways (SBB), and the other is derived from publicly available videos of arcing events in different railway systems and synthetic force data that mimic the characteristics observed in the real dataset. Leveraging these datasets, we propose MultiDeepSAD, an extension of the DeepSAD algorithm for multiple modalities with a new loss formulation. Additionally, we introduce tailored pseudo-anomaly generation techniques specific to each data type, such as synthetic arc-like artifacts in images and simulated force irregularities, to augment training data and improve the discriminative ability of the model. Through extensive experiments and ablation studies, we demonstrate that our framework significantly outperforms baseline approaches, exhibiting enhanced sensitivity to real arcing events even under domain shifts and limited availability of real arcing observations.
Authors: Jiaming Liu, Cheng Ding, Daoqiang Zhang
Abstract: Accurate brain tumor segmentation from MRI is limited by expensive annotations and data heterogeneity across scanners and sites. We propose a semi-supervised teacher-student framework that combines an uncertainty-aware pseudo-labeling teacher with a progressive, confidence-based curriculum for the student. The teacher produces probabilistic masks and per-pixel uncertainty; unlabeled scans are ranked by image-level confidence and introduced in stages, while a dual-loss objective trains the student to learn from high-confidence regions and unlearn low-confidence ones. Agreement-based refinement further improves pseudo-label quality. On BraTS 2021, validation DSC increased from 0.393 (10% data) to 0.872 (100%), with the largest gains in early stages, demonstrating data efficiency. The teacher reached a validation DSC of 0.922, and the student surpassed the teacher on tumor subregions (e.g., NCR/NET 0.797 and Edema 0.980); notably, the student recovered the Enhancing class (DSC 0.620) where the teacher failed. These results show that confidence-driven curricula and selective unlearning provide robust segmentation under limited supervision and noisy pseudo-labels.
Authors: Karan Bania, Soham Kalburgi, Manit Tanwar, Dhruthi, Aditya Nagarsekar, Harshvardhan Mestha, Naman Chibber, Raj Deshmukh, Anish Sathyanarayanan, Aarush Rathore, Pratham Chheda
Abstract: Linear recurrent neural networks (LRNNs) provide a structured approach to sequence modeling that bridges classical linear dynamical systems and modern deep learning, offering both expressive power and theoretical guarantees on stability and trainability. In recent years, multiple LRNN-based architectures have been proposed, each introducing distinct parameterizations, discretization schemes, and implementation constraints. However, existing implementations are fragmented across different software frameworks, often rely on framework-specific optimizations, and in some cases require custom CUDA kernels or lack publicly available code altogether. As a result, using, comparing, or extending LRNNs requires substantial implementation effort. To address this, we introduce $\texttt{lrnnx}$, a unified software library that implements several modern LRNN architectures under a common interface. The library exposes multiple levels of control, allowing users to work directly with core components or higher-level model abstractions. $\texttt{lrnnx}$ aims to improve accessibility, reproducibility, and extensibility of LRNN research and applications. We make our code available under a permissive MIT license.
Authors: James Jewitt, Gopi Krishnan Rajbahadur, Hao Li, Bram Adams, Ahmed E. Hassan
Abstract: Permissive licenses like MIT, Apache-2.0, and BSD-3-Clause dominate open-source AI, signaling that artifacts like models, datasets, and code can be freely used, modified, and redistributed. However, these licenses carry mandatory requirements: include the full license text, provide a copyright notice, and preserve upstream attribution, that remain unverified at scale. Failure to meet these conditions can place reuse outside the scope of the license, effectively leaving AI artifacts under default copyright for those uses and exposing downstream users to litigation. We call this phenomenon ``permissive washing'': labeling AI artifacts as free to use, while omitting the legal documentation required to make that label actionable. To assess how widespread permissive washing is in the AI supply chain, we empirically audit 124,278 dataset $\rightarrow$ model $\rightarrow$ application supply chains, spanning 3,338 datasets, 6,664 models, and 28,516 applications across Hugging Face and GitHub. We find that an astonishing 96.5\% of datasets and 95.8\% of models lack the required license text, only 2.3\% of datasets and 3.2\% of models satisfy both license text and copyright requirements, and even when upstream artifacts provide complete licensing evidence, attribution rarely propagates downstream: only 27.59\% of models preserve compliant dataset notices and only 5.75\% of applications preserve compliant model notices (with just 6.38\% preserving any linked upstream notice). Practitioners cannot assume permissive labels confer the rights they claim: license files and notices, not metadata, are the source of legal truth. To support future research, we release our full audit dataset and reproducible pipeline.
Authors: Chenghui Zou, Ning Wang, Tiesunlong Shen, Luwei Xiao, Chuan Ma, Xiangpeng Li, Rui Mao, Erik Cambria
Abstract: Large language models (LLMs) have been widely applied to emotional support conversation (ESC). However, complex multi-turn support remains challenging.This is because existing alignment schemes rely on sparse outcome-level signals, thus offering limited supervision for intermediate strategy decisions. To fill this gap, this paper proposes affective flow language model for emotional support conversation (AFlow), a framework that introduces fine-grained supervision on dialogue prefixes by modeling a continuous affective flow along multi-turn trajectories. AFlow can estimate intermediate utility over searched trajectories and learn preference-consistent strategy transitions. To improve strategy coherence and empathetic response quality, a subpath-level flow-balance objective is presented to propagate preference signals to intermediate states. Experiment results show consistent and significant improvements over competitive baselines in diverse emotional contexts. Remarkably, AFlow with a compact open-source backbone outperforms proprietary LMMs such as GPT-4o and Claude-3.5 on major ESC metrics. Our code is available at https://github.com/chzou25-lgtm/AffectiveFlow.
Authors: Hao Peng, Yunjia Qi, Xiaozhi Wang, Zijun Yao, Lei Hou, Juanzi Li
Abstract: Reward models (RMs) are crucial for the training of large language models (LLMs), yet they typically rely on large-scale human-annotated preference pairs. With the widespread deployment of LLMs, in-the-wild interactions have emerged as a rich source of implicit reward signals. This raises the question: Can we develop reward models directly from in-the-wild interactions? In this work, we explore this possibility by adopting WildChat as an interaction source and proposing a pipeline to extract reliable human feedback, yielding 186k high-quality instances for training WildReward via ordinal regression directly on user feedback without preference pairs. Extensive experiments demonstrate that WildReward achieves comparable or even superior performance compared to conventional reward models, with improved calibration and cross-sample consistency. We also observe that WildReward benefits directly from user diversity, where more users yield stronger reward models. Finally, we apply WildReward to online DPO training and observe significant improvements across various tasks. Code and data are released at https://github.com/THU-KEG/WildReward.
Authors: Lang Feng, Longtao Zheng, Shuo He, Fuxiang Zhang, Bo An
Abstract: Multi-agent LLM systems enable advanced reasoning and tool use via role specialization, yet reliable reinforcement learning (RL) post-training for such systems remains difficult. In this work, we theoretically pinpoint a key reason for training instability when extending group-based RL to multi-agent LLM systems. We show that under GRPO-style optimization, a global normalization baseline may deviate from diverse agents' reward distributions, which ultimately leads to gradient-norm instability. Based on this finding, we propose Dr. MAS, a simple and stable RL training recipe for multi-agent LLM systems. Dr. MAS uses an agent-wise remedy: normalizing advantages per agent using each agent's own reward statistics, which calibrates gradient scales and dramatically stabilizes training, both theoretically and empirically. Beyond the algorithm, Dr. MAS provides an end-to-end RL training framework for multi-agent LLM systems, supporting scalable orchestration, flexible per-agent LLM serving and optimization configs, and shared resource scheduling of LLM actor backends. We evaluate Dr. MAS on multi-agent math reasoning and multi-turn search benchmarks using Qwen2.5 and Qwen3 series models. Dr. MAS achieves clear gains over vanilla GRPO (e.g., +5.6\% avg@16 and +4.6\% pass@16 on math, and +15.2\% avg@16 and +13.1\% pass@16 on search) while largely eliminating gradient spikes. Moreover, it remains highly effective under heterogeneous agent-model assignments while improving efficiency.
Authors: Xinting Huang, Aleksandra Bakalova, Satwik Bhattamishra, William Merrill, Michael Hahn
Abstract: Recent work has shown that the computations of Transformers can be simulated in the RASP family of programming languages. These findings have enabled improved understanding of the expressive capacity and generalization abilities of Transformers. In particular, Transformers have been suggested to length-generalize exactly on problems that have simple RASP programs. However, it remains open whether trained models actually implement simple interpretable programs. In this paper, we present a general method to extract such programs from trained Transformers. The idea is to faithfully re-parameterize a Transformer as a RASP program and then apply causal interventions to discover a small sufficient sub-program. In experiments on small Transformers trained on algorithmic and formal language tasks, we show that our method often recovers simple and interpretable RASP programs from length-generalizing transformers. Our results provide the most direct evidence so far that Transformers internally implement simple RASP programs.
Authors: Ruihan Xu, Qingpei Guo, Yao Zhu, Xiangyang Ji, Ming Yang, Shiliang Zhang
Abstract: Recent works have indicated redundancy across transformer blocks, prompting the research of depth compression to prune less crucial blocks. However, current ways of entire-block pruning suffer from risks of discarding meaningful cues learned in those blocks, leading to substantial performance degradation. As another line of model compression, channel pruning can better preserve performance, while it cannot reduce model depth and is challenged by inconsistent pruning ratios for individual layers. To pursue better model compression and acceleration, this paper proposes \textbf{FlattenGPT}, a novel way to detect and reduce depth-wise redundancies. By flatting two adjacent blocks into one, it compresses the network depth, meanwhile enables more effective parameter redundancy detection and removal. FlattenGPT allows to preserve the knowledge learned in all blocks, and remains consistent with the original transformer architecture. Extensive experiments demonstrate that FlattenGPT enhances model efficiency with a decent trade-off to performance. It outperforms existing pruning methods in both zero-shot accuracies and WikiText-2 perplexity across various model types and parameter sizes. On LLaMA-2/3 and Qwen-1.5 models, FlattenGPT retains 90-96\% of zero-shot performance with a compression ratio of 20\%. It also outperforms other pruning methods in accelerating LLM inference, making it promising for enhancing the efficiency of transformers.
Authors: Ibraheem Muhammad Moosa, Suhas Lohit, Ye Wang, Moitreya Chatterjee, Wenpeng Yin
Abstract: Token-level adaptive computation seeks to reduce inference cost by allocating more computation to harder tokens and less to easier ones. However, prior work is primarily evaluated on natural-language benchmarks using task-level metrics, where token-level difficulty is unobservable and confounded with architectural factors, making it unclear whether compute allocation truly aligns with underlying complexity. We address this gap through three contributions. First, we introduce a complexity-controlled evaluation paradigm using algorithmic and synthetic language tasks with parameterized difficulty, enabling direct testing of token-level compute allocation. Second, we propose ANIRA, a unified recurrent Transformer framework that supports per-token variable-depth computation while isolating compute allocation decisions from other model factors. Third, we use this framework to conduct a systematic analysis of token-level adaptive computation across alignment with complexity, generalization, and decision timing. Our results show that compute allocation aligned with task complexity can emerge without explicit difficulty supervision, but such alignment does not imply algorithmic generalization: models fail to extrapolate to unseen input sizes despite allocating additional computation. We further find that early compute decisions rely on static structural cues, whereas online halting more closely tracks algorithmic execution state.
Authors: Junru Zhang, Lang Feng, Haoran Shi, Xu Guo, Han Yu, Yabo Dong, Duanqing Xu
Abstract: Time-series anomaly detection (TSAD) with multimodal large language models (MLLMs) is an emerging area, yet a persistent challenge remains: MLLMs rely on coarse time-series heuristics but struggle with multi-dimensional, detailed reasoning, which is vital for understanding complex time-series data. We present AnomSeer to address this by reinforcing the model to ground its reasoning in precise, structural details of time series, unifying anomaly classification, localization, and explanation. At its core, an expert chain-of-thought trace is generated to provide a verifiable, fine-grained reasoning from classical analyses (e.g., statistical measures, frequency transforms). Building on this, we propose a novel time-series grounded policy optimization (TimerPO) that incorporates two additional components beyond standard reinforcement learning: a time-series grounded advantage based on optimal transport and an orthogonal projection to ensure this auxiliary granular signal does not interfere with the primary detection objective. Across diverse anomaly scenarios, AnomSeer, with Qwen2.5-VL-3B/7B-Instruct, outperforms larger commercial baselines (e.g., GPT-4o) in classification and localization accuracy, particularly on point- and frequency-driven exceptions. Moreover, it produces plausible time-series reasoning traces that support its conclusions.
Authors: Lisette Espin-Noboa, Gonzalo Gabriel Mendez
Abstract: Large language models (LLMs) are increasingly used for academic expert recommendation. Existing audits typically evaluate model outputs in isolation, largely ignoring end-user inference-time interventions. As a result, it remains unclear whether failures such as refusals, hallucinations, and uneven coverage stem from model choice or deployment decisions. We introduce LLMScholarBench, a benchmark for auditing LLM-based scholar recommendation that jointly evaluates model infrastructure and end-user interventions across multiple tasks. LLMScholarBench measures both technical quality and social representation using nine metrics. We instantiate the benchmark in physics expert recommendation and audit 22 LLMs under temperature variation, representation-constrained prompting, and retrieval-augmented generation (RAG) via web search. Our results show that end-user interventions do not yield uniform improvements but instead redistribute error across dimensions. Higher temperature degrades validity, consistency, and factuality. Representation-constrained prompting improves diversity at the expense of factuality, while RAG primarily improves technical quality while reducing diversity and parity. Overall, end-user interventions reshape trade-offs rather than providing a general fix. We release code and data that can be adapted to other disciplines by replacing domain-specific ground truth and metrics.
Authors: Itai Zilberstein, Ioannis Anagnostides, Zachary W. Sollie, Arman Kilic, Tuomas Sandholm
Abstract: Each year, thousands of patients in need of heart transplants face life-threatening wait times due to organ scarcity. While allocation policies aim to maximize population-level outcomes, current approaches often fail to account for the dynamic arrival of organs and the composition of waitlisted candidates, thereby hampering efficiency. The United States is transitioning from rigid, rule-based allocation to more flexible data-driven models. In this paper, we propose a novel framework for non-myopic policy optimization in general online matching relying on potentials, a concept originally introduced for kidney exchange. We develop scalable and accurate ways of learning potentials that are higher-dimensional and more expressive than prior approaches. Our approach is a form of self-supervised imitation learning: the potentials are trained to mimic an omniscient algorithm that has perfect foresight. We focus on the application of heart transplant allocation and demonstrate, using real historical data, that our policies significantly outperform prior approaches -- including the current US status quo policy and the proposed continuous distribution framework -- in optimizing for population-level outcomes. Our analysis and methods come at a pivotal moment in US policy, as the current heart transplant allocation system is under review. We propose a scalable and theoretically grounded path toward more effective organ allocation.
Authors: Paul Saegert, Ullrich K\"othe
Abstract: Symbolic regression (SR) aims to discover interpretable analytical expressions that accurately describe observed data. Amortized SR promises to be much more efficient than the predominant genetic programming SR methods, but currently struggles to scale to realistic scientific complexity. We find that a key obstacle is the lack of a fast reduction of equivalent expressions to a concise normalized form. Amortized SR has addressed this by general-purpose Computer Algebra Systems (CAS) like SymPy, but the high computational cost severely limits training and inference speed. We propose SimpliPy, a rule-based simplification engine achieving a 100-fold speed-up over SymPy at comparable quality. This enables substantial improvements in amortized SR, including scalability to much larger training sets, more efficient use of the per-expression token budget, and systematic training set decontamination with respect to equivalent test expressions. We demonstrate these advantages in our Flash-ANSR framework, which achieves much better accuracy than amortized baselines (NeSymReS, E2E) on the FastSRB benchmark. Moreover, it performs on par with state-of-the-art direct optimization (PySR) while recovering more concise instead of more complex expressions with increasing inference budget.
Authors: Adam Trendowicz, Daniel Seifert, Andreas Jedlitschka, Marcus Ciolkowski, Anton Strahilov
Abstract: Generative artificial intelligence (GAI), specifically large language models (LLMs), are increasingly used in software engineering, mainly for coding tasks. However, requirements engineering - particularly requirements validation - has seen limited application of GAI. The current focus of using GAI for requirements is on eliciting, transforming, and classifying requirements, not on quality assessment. We propose and evaluate the LLM-based (GPT-4o) approach "DeepQuali", for assessing and improving requirements quality in agile software development. We applied it to projects in two small companies, where we compared LLM-based quality assessments with expert judgments. Experts also participated in walkthroughs of the solution, provided feedback, and rated their acceptance of the approach. Experts largely agreed with the LLM's quality assessments, especially regarding overall ratings and explanations. However, they did not always agree with the other experts on detailed ratings, suggesting that expertise and experience may influence judgments. Experts recognized the usefulness of the approach but criticized the lack of integration into their workflow. LLMs show potential in supporting software engineers with the quality assessment and improvement of requirements. The explicit use of quality models and explanatory feedback increases acceptance.
Authors: Yehua Huang, Penglei Sun, Zebin Chen, Zhenheng Tang, Xiaowen Chu
Abstract: Academic peer review remains the cornerstone of scholarly validation, yet the field faces some challenges in data and methods. From the data perspective, existing research is hindered by the scarcity of large-scale, verified benchmarks and oversimplified evaluation metrics that fail to reflect real-world editorial workflows. To bridge this gap, we present OmniReview, a comprehensive dataset constructed by integrating multi-source academic platforms encompassing comprehensive scholarly profiles through the disambiguation pipeline, yielding 202, 756 verified review records. Based on this data, we introduce a three-tier hierarchical evaluaion framework to assess recommendations from recall to precise expert identification. From the method perspective, existing embedding-based approaches suffer from the information bottleneck of semantic compression and limited interpretability. To resolve these method limitations, we propose Profiling Scholars with Multi-gate Mixture-of-Experts (Pro-MMoE), a novel framework that synergizes Large Language Models (LLMs) with Multi-task Learning. Specifically, it utilizes LLM-generated semantic profiles to preserve fine-grained expertise nuances and interpretability, while employing a Task-Adaptive MMoE architecture to dynamically balance conflicting evaluation goals. Comprehensive experiments demonstrate that Pro-MMoE achieves state-of-the-art performance across six of seven metrics, establishing a new benchmark for realistic reviewer recommendation.
Authors: Kiyosu Maeda, William P. McCarthy, Ching-Yi Tsai, Jeffrey Mu, Haoliang Wang, Robert D. Hawkins, Judith E. Fan, Parastoo Abtahi
Abstract: A quintessential feature of human intelligence is the ability to create ad hoc conventions over time to achieve shared goals efficiently. We investigate how communication strategies evolve through repeated collaboration as people coordinate on shared procedural abstractions. To this end, we conducted an online unimodal study (n = 98) using natural language to probe abstraction hierarchies. In a follow-up lab study (n = 40), we examined how multimodal communication (speech and gestures) changed during physical collaboration. Pairs used augmented reality to isolate their partner's hand and voice; one participant viewed a 3D virtual tower and sent instructions to the other, who built the physical tower. Participants became faster and more accurate by establishing linguistic and gestural abstractions and using cross-modal redundancy to emphasize key changes from previous interactions. Based on these findings, we extend probabilistic models of convention formation to multimodal settings, capturing shifts in modality preferences. Our findings and model provide building blocks for designing convention-aware intelligent agents situated in the physical world.
Authors: Minghan Li, Ercong Nie, Siqi Zhao, Tongna Chen, Huiping Huang, Guodong Zhou
Abstract: Query expansion with large language models is promising but often relies on hand-crafted prompts, manually chosen exemplars, or a single LLM, making it non-scalable and sensitive to domain shift. We present an automated, domain-adaptive QE framework that builds in-domain exemplar pools by harvesting pseudo-relevant passages using a BM25-MonoT5 pipeline. A training-free cluster-based strategy selects diverse demonstrations, yielding strong and stable in-context QE without supervision. To further exploit model complementarity, we introduce a two-LLM ensemble in which two heterogeneous LLMs independently generate expansions and a refinement LLM consolidates them into one coherent expansion. Across TREC DL20, DBPedia, and SciFact, the refined ensemble delivers consistent and statistically significant gains over BM25, Rocchio, zero-shot, and fixed few-shot baselines. The framework offers a reproducible testbed for exemplar selection and multi-LLM generation, and a practical, label-free solution for real-world QE.
Authors: Suraj Ranganath, Atharv Ramesh
Abstract: AI-text detectors face a critical robustness challenge: adversarial paraphrasing attacks that preserve semantics while evading detection. We introduce StealthRL, a reinforcement learning framework that stress-tests detector robustness under realistic adversarial conditions. StealthRL trains a paraphrase policy against a multi-detector ensemble using Group Relative Policy Optimization (GRPO) with LoRA adapters on Qwen3-4B, optimizing a composite reward that balances detector evasion with semantic preservation. We evaluate six attack settings (M0-M5) against three detector families (RoBERTa, FastDetectGPT, and Binoculars) at the security-relevant 1% false positive rate operating point. StealthRL achieves near-zero detection (0.001 mean TPR@1%FPR), reduces mean AUROC from 0.74 to 0.27, and attains a 99.9% attack success rate. Critically, attacks transfer to a held-out detector family not seen during training, revealing shared architectural vulnerabilities rather than detector-specific brittleness. We additionally conduct LLM-based quality evaluation via Likert scoring, analyze detector score distributions to explain why evasion succeeds, and provide per-detector AUROC with bootstrap confidence intervals. Our results expose significant robustness gaps in current AI-text detection and establish StealthRL as a principled adversarial evaluation protocol. Code and evaluation pipeline are publicly available at https://github.com/suraj-ranganath/StealthRL.
Authors: Zeyu Lu, Dennis L. Barbour
Abstract: Traditional cognitive assessments often rely on isolated, output-focused measurements that may fail to capture the complexity of human cognition in naturalistic settings. We present pixelLOG, a high-performance data collection framework for Spigot-based Minecraft servers designed specifically for process-based cognitive research. Unlike existing frameworks tailored only for artificial intelligence agents, pixelLOG also enables human behavioral tracking in multi-player/multi-agent environments. Operating at configurable frequencies up to and exceeding 20 updates per second, the system captures comprehensive behavioral data through a hybrid approach of active state polling and passive event monitoring. By leveraging Spigot's extensible API, pixelLOG facilitates robust session isolation and produces structured JSON outputs integrable with standard analytical pipelines. This framework bridges the gap between decontextualized laboratory assessments and richer, more ecologically valid tasks, enabling high-resolution analysis of cognitive processes as they unfold in complex, virtual environments.
Authors: Ruijie Zhu, Jiahao Lu, Wenbo Hu, Xiaoguang Han, Jianfei Cai, Ying Shan, Chuanxia Zheng
Abstract: We introduce MotionCrafter, a video diffusion-based framework that jointly reconstructs 4D geometry and estimates dense motion from a monocular video. The core of our method is a novel joint representation of dense 3D point maps and 3D scene flows in a shared coordinate system, and a novel 4D VAE to effectively learn this representation. Unlike prior work that forces the 3D value and latents to align strictly with RGB VAE latents-despite their fundamentally different distributions-we show that such alignment is unnecessary and leads to suboptimal performance. Instead, we introduce a new data normalization and VAE training strategy that better transfers diffusion priors and greatly improves reconstruction quality. Extensive experiments across multiple datasets demonstrate that MotionCrafter achieves state-of-the-art performance in both geometry reconstruction and dense scene flow estimation, delivering 38.64% and 25.0% improvements in geometry and motion reconstruction, respectively, all without any post-optimization. Project page: https://ruijiezhu94.github.io/MotionCrafter_Page
Authors: Raghu Arghal, Fade Chen, Niall Dalton, Evgenii Kortukov, Calum McNamara, Angelos Nalmpantis, Moksh Nirvaan, Gabriele Sarti, Mario Giulianelli
Abstract: Understanding an agent's goals helps explain and predict its behaviour, yet there is no established methodology for reliably attributing goals to agentic systems. We propose a framework for evaluating goal-directedness that integrates behavioural evaluation with interpretability-based analyses of models' internal representations. As a case study, we examine an LLM agent navigating a 2D grid world toward a goal state. Behaviourally, we evaluate the agent against an optimal policy across varying grid sizes, obstacle densities, and goal structures, finding that performance scales with task difficulty while remaining robust to difficulty-preserving transformations and complex goal structures. We then use probing methods to decode the agent's internal representations of the environment state and its multi-step action plans. We find that the LLM agent non-linearly encodes a coarse spatial map of the environment, preserving approximate task-relevant cues about its position and the goal location; that its actions are broadly consistent with these internal representations; and that reasoning reorganises them, shifting from broader environment structural cues toward information supporting immediate action selection. Our findings support the view that introspective examination is required beyond behavioural evaluations to characterise how agents represent and pursue their objectives.
Authors: Yubin Kim, Viresh Pati, Jevon Twitty, Vinh Pham, Shihao Yang, Jiecheng Lu
Abstract: Transformer architectures have established strong baselines in time series forecasting, yet they typically rely on positional encodings that assume uniform, index-based temporal progression. However, real-world systems, from shifting financial cycles to elastic biological rhythms, frequently exhibit "time-warped" dynamics where the effective flow of time decouples from the sampling index. In this work, we first formalize this misalignment and prove that rotary position embedding (RoPE) is mathematically incapable of representing non-affine temporal warping. To address this, we propose Symplectic Positional Embeddings (SyPE), a learnable encoding framework derived from Hamiltonian mechanics. SyPE strictly generalizes RoPE by extending the rotation group $\mathrm{SO}(2)$ to the symplectic group $\mathrm{Sp}(2,\mathbb{R})$, modulated by a novel input-dependent adaptive warp module. By allowing the attention mechanism to adaptively dilate or contract temporal coordinates end-to-end, our approach captures locally varying periodicities without requiring pre-defined warping functions. We implement this mechanism in StretchTime, a multivariate forecasting architecture that achieves state-of-the-art performance on standard benchmarks, demonstrating superior robustness on datasets exhibiting non-stationary temporal dynamics.
Authors: Yuliang Liu, Yunchong Song, Yixuan Wang, Kewen Ge, Alex Lamb, Qipeng Guo, Kai Chen, Bowen Zhou, Zhouhan Lin
Abstract: We propose Next Concept Prediction (NCP), a generative pretraining paradigm built on top of Next Token Prediction (NTP). NCP predicts discrete concepts that span multiple tokens, thereby forming a more challenging pretraining objective. Our model, ConceptLM, quantizes hidden states using Vector Quantization and constructs a concept vocabulary. It leverages both NCP and NTP to drive parameter updates and generates a concept to guide the generation of the following tokens. We train ConceptLM from scratch at scales ranging from 70M to 1.5B parameters with up to 300B training data, including Pythia and GPT-2 backbones. Results on 13 benchmarks show that NCP yields consistent performance gains over traditional token-level models. Furthermore, continual pretraining experiments on an 8B-parameter Llama model indicate that NCP can further improve an NTP-trained model. Our analysis suggests that NCP leads to more powerful language models by introducing a harder pretraining task, providing a promising path toward better language modeling.
Authors: Isaac Xu, Martin Gillis, Ayushi Sharma, Benjamin Misiuk, Craig J. Brown, Thomas Trappenberg
Abstract: In hierarchical multi-label classification, a persistent challenge is enabling model predictions to reach deeper levels of the hierarchy for more detailed or fine-grained classifications. This difficulty partly arises from the natural rarity of certain classes (or hierarchical nodes) and the hierarchical constraint that ensures child nodes are almost always less frequent than their parents. To address this, we propose a weighted loss objective for neural networks that combines node-wise imbalance weighting with focal weighting components, the latter leveraging modern quantification of ensemble uncertainties. By emphasizing rare nodes rather than rare observations (data points), and focusing on uncertain nodes for each model output distribution during training, we observe improvements in recall by up to a factor of five on benchmark datasets, along with statistically significant gains in $F_{1}$ score. We also show our approach aids convolutional networks on challenging tasks, as in situations with suboptimal encoders or limited data.
Authors: Zilin Fang, Anxing Xiao, David Hsu, Gim Hee Lee
Abstract: Navigating socially in human environments requires more than satisfying geometric constraints, as collision-free paths may still interfere with ongoing activities or conflict with social norms. Addressing this challenge calls for analyzing interactions between agents and incorporating common-sense reasoning into planning. This paper presents a social robot navigation framework that integrates geometric planning with contextual social reasoning. The system first extracts obstacles and human dynamics to generate geometrically feasible candidate paths, then leverages a fine-tuned vision-language model (VLM) to evaluate these paths, informed by contextually grounded social expectations, selecting a socially optimized path for the controller. This task-specific VLM distills social reasoning from large foundation models into a smaller and efficient model, allowing the framework to perform real-time adaptation in diverse human-robot interaction contexts. Experiments in four social navigation contexts demonstrate that our method achieves the best overall performance with the lowest personal space violation duration, the minimal pedestrian-facing time, and no social zone intrusions. Project page: https://path-etiquette.github.io
Authors: Wenbo Gong, Javier Zazo, Qijun Luo, Puqian Wang, James Hensman, Chao Ma
Abstract: Matrix-based optimizers have attracted growing interest for improving LLM training efficiency, with significant progress centered on orthogonalization/whitening based methods. While yielding substantial performance gains, a fundamental question arises: can we develop new paradigms beyond orthogonalization, pushing the efficiency frontier further? We present \textbf{Adaptively Rotated Optimization (ARO}, a new matrix optimization framework that treats gradient rotation as a first class design principle. ARO accelerates LLM training by performing normed steepest descent in a rotated coordinate system, where the rotation is determined by a novel norm-informed policy. This perspective yields update rules that go beyond existing orthogonalization and whitening optimizers, improving sample efficiency in practice. To make comparisons reliable, we propose a rigorously controlled benchmarking protocol that reduces confounding and bias. Under this protocol, ARO consistently outperforms AdamW (by 1.3 $\sim$1.35$\times$) and orthogonalization methods (by 1.1$\sim$1.15$\times$) in LLM pretraining at up to 8B activated parameters, and up to $8\times$ overtrain budget, without evidence of diminishing returns. Finally, we discuss how ARO can be reformulated as a symmetry-aware optimizer grounded in rotational symmetries of residual streams, motivating advanced designs that enable computationally efficient exploitation of cross-layer/cross module couplings.
Authors: Yilang Zhang, Bingcong Li, Niao He, Georgios B. Giannakis
Abstract: Scaling network depth has been a central driver behind the success of modern foundation models, yet recent investigations suggest that deep layers are often underutilized. This paper revisits the default mechanism for deepening neural networks, namely residual connections, from an optimization perspective. Rigorous analysis proves that the layout of residual connections can fundamentally shape convergence behavior, and even induces an exponential gap in convergence rates. Prompted by this insight, we introduce adaptive neural connection reassignment (ANCRe), a principled and lightweight framework that parameterizes and learns residual connectivities from the data. ANCRe adaptively reassigns residual connections with negligible computational and memory overhead ($<1\%$), while enabling more effective utilization of network depth. Extensive numerical tests across pre-training of large language models, diffusion models, and deep ResNets demonstrate consistently accelerated convergence, boosted performance, and enhanced depth efficiency over conventional residual connections.
Authors: Jiacheng Liu, Yaxin Luo, Jiacheng Cui, Xinyi Shang, Xiaohan Zhao, Zhiqiang Shen
Abstract: The rapid evolution of GUI-enabled agents has rendered traditional CAPTCHAs obsolete. While previous benchmarks like OpenCaptchaWorld established a baseline for evaluating multimodal agents, recent advancements in reasoning-heavy models, such as Gemini3-Pro-High and GPT-5.2-Xhigh have effectively collapsed this security barrier, achieving pass rates as high as 90% on complex logic puzzles like "Bingo". In response, we introduce Next-Gen CAPTCHAs, a scalable defense framework designed to secure the next-generation web against the advanced agents. Unlike static datasets, our benchmark is built upon a robust data generation pipeline, allowing for large-scale and easily scalable evaluations, notably, for backend-supported types, our system is capable of generating effectively unbounded CAPTCHA instances. We exploit the persistent human-agent "Cognitive Gap" in interactive perception, memory, decision-making, and action. By engineering dynamic tasks that require adaptive intuition rather than granular planning, we re-establish a robust distinction between biological users and artificial agents, offering a scalable and diverse defense mechanism for the agentic era.
Authors: Zihan Yang (Fudan University), Shuyuan Tu (Fudan University), Licheng Zhang (Fudan University), Qi Dai (Microsoft Research Asia), Yu-Gang Jiang (Fudan University), Zuxuan Wu (Fudan University)
Abstract: Diffusion models have achieved remarkable generation quality, but they suffer from significant inference cost due to their reliance on multiple sequential denoising steps, motivating recent efforts to distill this inference process into a few-step regime. However, existing distillation methods typically approximate the teacher trajectory by using linear shortcuts, which makes it difficult to match its constantly changing tangent directions as velocities evolve across timesteps, thereby leading to quality degradation. To address this limitation, we propose ArcFlow, a few-step distillation framework that explicitly employs non-linear flow trajectories to approximate pre-trained teacher trajectories. Concretely, ArcFlow parameterizes the velocity field underlying the inference trajectory as a mixture of continuous momentum processes. This enables ArcFlow to capture velocity evolution and extrapolate coherent velocities to form a continuous non-linear trajectory within each denoising step. Importantly, this parameterization admits an analytical integration of this non-linear trajectory, which circumvents numerical discretization errors and results in high-precision approximation of the teacher trajectory. To train this parameterization into a few-step generator, we implement ArcFlow via trajectory distillation on pre-trained teacher models using lightweight adapters. This strategy ensures fast, stable convergence while preserving generative diversity and quality. Built on large-scale models (Qwen-Image-20B and FLUX.1-dev), ArcFlow only fine-tunes on less than 5% of original parameters and achieves a 40x speedup with 2 NFEs over the original multi-step teachers without significant quality degradation. Experiments on benchmarks show the effectiveness of ArcFlow both qualitatively and quantitatively.
Authors: Fatemeh Nejati, Mahdi Rabbani, Mansur Mirani, Gunjan Piya, Igor Opushnyev, Ali A. Ghorbani, Sajjad Dadkhah
Abstract: Phishing attacks represents one of the primary attack methods which is used by cyber attackers. In many cases, attackers use deceptive emails along with malicious attachments to trick users into giving away sensitive information or installing malware while compromising entire systems. The flexibility of malicious email attachments makes them stand out as a preferred vector for attackers as they can embed harmful content such as malware or malicious URLs inside standard document formats. Although phishing email defenses have improved a lot, attackers continue to abuse attachments, enabling malicious content to bypass security measures. Moreover, another challenge that researches face in training advance models, is lack of an unified and comprehensive dataset that covers the most prevalent data types. To address this gap, we generated CIC-Trap4Phish, a multi-format dataset containing both malicious and benign samples across five categories commonly used in phishing campaigns: Microsoft Word documents, Excel spreadsheets, PDF files, HTML pages, and QR code images. For the first four file types, a set of execution-free static feature pipeline was proposed, designed to capture structural, lexical, and metadata-based indicators without the need to open or execute files. Feature selection was performed using a combination of SHAP analysis and feature importance, yielding compact, discriminative feature subsets for each file type. The selected features were evaluated by using lightweight machine learning models, including Random Forest, XGBoost, and Decision Tree. All models demonstrate high detection accuracy across formats. For QR code-based phishing (quishing), two complementary methods were implemented: image-based detection by employing Convolutional Neural Networks (CNNs) and lexical analysis of decoded URLs using recent lightweight language models.
Authors: Amir Mallak, Alaa Maalouf
Abstract: Out of distribution (OOD) robustness in autonomous driving is often reduced to a single number, hiding what breaks a policy. We decompose environments along five axes: scene (rural/urban), season, weather, time (day/night), and agent mix; and measure performance under controlled $k$-factor perturbations ($k \in \{0,1,2,3\}$). Using closed loop control in VISTA, we benchmark FC, CNN, and ViT policies, train compact ViT heads on frozen foundation-model (FM) features, and vary ID support in scale, diversity, and temporal context. (1) ViT policies are markedly more OOD-robust than comparably sized CNN/FC, and FM features yield state-of-the-art success at a latency cost. (2) Naive temporal inputs (multi-frame) do not beat the best single-frame baseline. (3) The largest single factor drops are rural $\rightarrow$ urban and day $\rightarrow$ night ($\sim 31\%$ each); actor swaps $\sim 10\%$, moderate rain $\sim 7\%$; season shifts can be drastic, and combining a time flip with other changes further degrades performance. (4) FM-feature policies stay above $85\%$ under three simultaneous changes; non-FM single-frame policies take a large first-shift hit, and all no-FM models fall below $50\%$ by three changes. (5) Interactions are non-additive: some pairings partially offset, whereas season-time combinations are especially harmful. (6) Training on winter/snow is most robust to single-factor shifts, while a rural+summer baseline gives the best overall OOD performance. (7) Scaling traces/views improves robustness ($+11.8$ points from $5$ to $14$ traces), yet targeted exposure to hard conditions can substitute for scale. (8) Using multiple ID environments broadens coverage and strengthens weak cases (urban OOD $60.6\% \rightarrow 70.1\%$) with a small ID drop; single-ID preserves peak performance but in a narrow domain. These results yield actionable design rules for OOD-robust driving policies.
Authors: Arun Kumar, Paul Schrater
Abstract: Interactions are central to intelligent reasoning and learning abilities, with the interpretation of abstract knowledge guiding meaningful interaction with objects in the environment. While humans readily adapt to novel situations by leveraging abstract knowledge acquired over time, artificial intelligence systems lack principled mechanisms for incorporating abstract knowledge into learning, leading to fundamental challenges in the emergence of intelligent and adaptive behavior. To address this gap, we introduce knowledge-centric metacognitive learning based on three key principles: natural abstractions, knowledge-guided interactions through interpretation, and the composition of interactions for problem solving. Knowledge learning facilitates the acquisition of abstract knowledge and the association of interactions with knowledge, while object interactions guided by abstract knowledge enable the learning of transferable interaction concepts, abstract reasoning, and generalization. This metacognitive mechanism provides a principled approach for integrating knowledge into reinforcement learning and offers a promising pathway toward intelligent and adaptive behavior in artificial intelligence, robotics, and autonomous systems.
Authors: Srijoni Majumdar, Edith Elkind, Evangelos Pournaras
Abstract: Recent breakthroughs in generative artificial intelligence (AI) and large language models (LLMs) unravel new capabilities for AI personal assistants to overcome cognitive bandwidth limitations of humans, providing decision support or even direct representation of abstained human voters at large scale. However, the quality of this representation and what underlying biases manifest when delegating collective decision making to LLMs is an alarming and timely challenge to tackle. By rigorously emulating more than >50K LLM voting personas in 363 real-world voting elections, we disentangle how AI-generated choices differ from human choices and how this affects collective decision outcomes. Complex preferential ballot formats show significant inconsistencies compared to simpler majoritarian elections, which demonstrate higher consistency. Strikingly, proportional ballot aggregation methods such as equal shares prove to be a win-win: fairer voting outcomes for humans and fairer AI representation, especially for voters likely to abstain. This novel underlying relationship proves paramount for building democratic resilience in scenarios of low voters turnout by voter fatigue: abstained voters are mitigated via AI representatives that recover representative and fair voting outcomes. These interdisciplinary insights provide decision support to policymakers and citizens for developing safeguards and policies for risks of using AI in democratic innovations.
Authors: Jing Yu Koh, Stephen McAleer, Daniel Fried, Ruslan Salakhutdinov
Abstract: Autonomous agents powered by language models (LMs) have demonstrated promise in their ability to perform decision-making tasks such as web automation. However, a key limitation remains: LMs, primarily optimized for natural language understanding and generation, struggle with multi-step reasoning, planning, and using environmental feedback when attempting to solve realistic computer tasks. Towards addressing this, we propose an inference-time search algorithm for LM agents to explicitly perform exploration and multi-step planning in interactive web environments. Our approach is a form of best-first tree search that operates within the actual environment space, and is complementary with most existing state-of-the-art agents. It is the first tree search algorithm for LM agents that shows effectiveness on realistic web tasks. On the challenging VisualWebArena benchmark, applying our search algorithm on top of a GPT-4o agent yields a 39.7% relative increase in success rate compared to the same baseline without search, setting a state-of-the-art success rate of 26.4%. On WebArena, search also yields a 28.0% relative improvement over a baseline agent, setting a competitive success rate of 19.2%. Our experiments highlight the effectiveness of search for web agents, and we demonstrate that performance scales with increased test-time compute. We conduct a thorough analysis of our results to highlight improvements from search, limitations, and promising directions for future work. Our code and models are publicly released at https://jykoh.com/search-agents.
Authors: Soumi Das, Camila Kolling, Mohammad Aflah Khan, Mahsa Amani, Bishwamittra Ghosh, Qinyuan Wu, Till Speicher, Krishna P. Gummadi
Abstract: We study the inherent trade-offs in minimizing privacy risks and maximizing utility, while maintaining high computational efficiency, when fine-tuning large language models (LLMs). A number of recent works in privacy research have attempted to mitigate privacy risks posed by memorizing fine-tuning data by using differentially private training methods (e.g., DP), albeit at a significantly higher computational cost (inefficiency). In parallel, several works in systems research have focussed on developing (parameter) efficient fine-tuning methods (e.g., LoRA), but few works, if any, investigated whether such efficient methods enhance or diminish privacy risks. In this paper, we investigate this gap and arrive at a surprising conclusion: efficient fine-tuning methods like LoRA mitigate privacy risks similar to private fine-tuning methods like DP. Our empirical finding directly contradicts prevailing wisdom that privacy and efficiency objectives are at odds during fine-tuning. Our finding is established by (a) carefully defining measures of privacy and utility that distinguish between memorizing sensitive and non-sensitive tokens in training and test datasets used in fine-tuning and (b) extensive evaluations using multiple open-source language models from Pythia, Gemma, Llama, and Qwen families and different domain-specific datasets.
Authors: HyunJin Kim, Xiaoyuan Yi, Jing Yao, Muhua Huang, JinYeong Bak, James Evans, Xing Xie
Abstract: The recent leap in AI capabilities, driven by big generative models, has sparked the possibility of achieving Artificial General Intelligence (AGI) and further triggered discussions on Artificial Superintelligence (ASI)-a system surpassing all humans across measured domains. This gives rise to the critical research question of: As we approach ASI, how do we align it with human values, ensuring it benefits rather than harms human society, a.k.a., the Superalignment problem. Despite ASI being regarded by many as a hypothetical concept, in this position paper, we argue that superalignment is achievable and research on it should advance immediately, through simultaneous and alternating optimization of task competence and value conformity. We posit that superalignment is not merely a safeguard for ASI but also necessary for its responsible realization. To support this position, we first provide a formal definition of superalignment rooted in the gap between capability and capacity, delve into its perceived infeasibility by analyzing the limitations of existing paradigms, and then illustrate a conceptual path of superalignment to support its achievability, centered on two fundamental principles. This work frames a potential initiative for developing value-aligned next-generation AI in the future, which will garner greater benefits and reduce potential harm to humanity.
Authors: Chang Yang, Ruiyu Wang, Junzhe Jiang, Qi Jiang, Qinggang Zhang, Yanchen Deng, Shuxin Li, Shuyue Hu, Bo Li, Florian T. Pokorny, Xiao Huang, Xinrun Wang
Abstract: Reasoning is the fundamental capability of large language models (LLMs). Due to the rapid progress of LLMs, there are two main issues of current benchmarks: i) these benchmarks can be crushed in a short time (less than 1 year), and ii) these benchmarks may be easily hacked. To handle these issues, we propose the ever-scalingness for building the benchmarks which are scaling over complexity against crushing, instance against hacking and exploitation, oversight for easy verification, and coverage for real-world relevance. This paper presents Nondeterministic Polynomial-time Problem Challenge (NPPC), an ever-scaling reasoning benchmark for LLMs. Specifically, the NPPC has three main modules: i) npgym, which provides a unified interface of 25 well-known NP-complete problems and can generate any number of instances with any levels of complexities, ii) npsolver, which provides a unified interface to evaluate the problem instances with both online and offline models via APIs and local deployments, respectively, and iii) npeval, which provides the comprehensive and ready-to-use tools to analyze the performances of LLMs over different problems, the number of tokens, the reasoning errors and the solution errors. Extensive experiments over widely-used LLMs demonstrate: i) NPPC can successfully decrease the performances of advanced LLMs to below 10%, demonstrating that NPPC is not crushed by current models, ii) DeepSeek-R1, Claude-3.7-Sonnet, and o1/o3-mini are the most powerful LLMs, and iii) the numbers of tokens in the advanced LLMs, e.g., Claude-3.7-Sonnet and DeepSeek-R1, are observed first to increase and then decrease when the problem instances become more and more difficult. Through continuously scaling analysis, NPPC can provide critical insights into the limits of LLMs' reasoning capabilities, exposing fundamental limitations and suggesting future directions for further improvements.
Authors: Haiyang Shen, Hang Yan, Zhongshi Xing, Mugeng Liu, Yue Li, Zhiyang Chen, Yuxiang Wang, Jiuzheng Wang, Yun Ma
Abstract: Retrieval-augmented generation (RAG) can substantially enhance the performance of LLMs on knowledge-intensive tasks. Various RAG paradigms - including vanilla, planning-based, and iterative RAG - all depend on a robust retriever, yet existing retrievers rely heavily on public knowledge and often falter when faced with domain-specific queries. To address these limitations, we introduce DRAGON, a framework that combines a data-construction modeling approach with a scalable synthetic data-generation pipeline, specifically designed to optimize domain-specific retrieval performance and bolster retriever robustness. To evaluate RAG performance on domain-specific RAGs, we propose DRAGONBench, a benchmark spanning 8 domain-specific document collections across 4 distinct fields and featuring a wide spectrum of query complexities, answerability, and hop numbers. Leveraging DRAGON, we generate a large-scale synthetic dataset - encompassing both single-hop and multi-hop queries - to enrich retriever training. Extensive experiments demonstrate that retrievers trained on this data yield significant performance gains and exhibit strong cross-domain generalization. Moreover, when our optimized retrievers are integrated into vanilla, planning-based, and iterative RAG paradigms, we observe consistent end-to-end improvements in system accuracy.
Authors: Milad Kazemi, Mateo Perez, Fabio Somenzi, Sadegh Soudjani, Ashutosh Trivedi, Alvaro Velasquez
Abstract: Recent advances in reinforcement learning (RL) have renewed interest in reward design for shaping agent behavior, but manually crafting reward functions is tedious and error-prone. A principled alternative is to specify behavioral requirements in a formal, unambiguous language and automatically compile them into learning objectives. $\omega$-regular languages are a natural fit, given their role in formal verification and synthesis. However, most existing $\omega$-regular RL approaches operate in an episodic, discounted setting with periodic resets, which is misaligned with $\omega$-regular semantics over infinite traces. For continuing tasks, where the agent interacts with the environment over a single uninterrupted lifetime, the average-reward criterion is more appropriate. We focus on absolute liveness specifications, a subclass of $\omega$-regular languages that cannot be violated by any finite prefix and thus aligns naturally with continuing interaction. We present the first model-free RL framework that translates absolute liveness specifications into average-reward objectives and enables learning in unknown communicating Markov decision processes (MDPs) without episodic resetting. We also introduce a reward structure for lexicographic multi-objective optimization: among policies that maximize the satisfaction probability of an absolute liveness specification, the agent maximizes an external average-reward objective. Our method guarantees convergence in unknown communicating MDPs and supports on-the-fly reductions that do not require full environment knowledge, enabling model-free learning. Experiments across several benchmarks show that the continuing, average-reward approach outperforms competing discount-based methods.
Authors: Alexander Panfilov, Paul Kassianik, Maksym Andriushchenko, Jonas Geiping
Abstract: As large language models grow in capability and agency, identifying vulnerabilities through red-teaming becomes vital for safe deployment. However, traditional prompt-engineering approaches may prove ineffective once red-teaming turns into a \emph{weak-to-strong} problem, where target models surpass red-teamers in capabilities. To study this shift, we frame red-teaming through the lens of the \emph{capability gap} between attacker and target. We evaluate more than 600 attacker-target pairs using LLM-based jailbreak attacks that mimic human red-teamers across diverse families, sizes, and capability levels. Three strong trends emerge: (i) more capable models are better attackers, (ii) attack success drops sharply once the target's capability exceeds the attacker's, and (iii) attack success rates correlate with high performance on social science splits of the MMLU-Pro benchmark. From these observations, we derive a \emph{jailbreaking scaling curve} that predicts attack success for a fixed target based on attacker-target capability gap. These findings suggest that fixed-capability attackers (e.g., humans) may become ineffective against future models, increasingly capable open-source models amplify risks for existing systems, and model providers must accurately measure and control models' persuasive and manipulative abilities to limit their effectiveness as attackers.
Authors: Shengkun Wang, Yanshen Sun, Fanglan Chen, Linhan Wang, Naren Ramakrishnan, Chang-Tien Lu, Yinlin Chen
Abstract: Accurate long-horizon house-price forecasting requires benchmarks that capture temporal dynamics together with time-varying local context. However, existing public resources remain fragmented: many datasets have limited spatial coverage, temporal depth, or multimodal alignment; the robustness of modern deep forecasters and time-series foundation models on housing data is not well characterized; and aerial imagery is rarely leveraged in a time-aware and interpretable manner at scale. To bridge these gaps, we present HouseTS (House Time Series), a multimodal spatiotemporal dataset for ZIP-code-level housing-market analysis, covering monthly signals from March 2012 to December 2023 across over 6,000 ZIP codes in 30 major U.S. metropolitan areas. HouseTS aligns monthly housing-market indicators, monthly POI dynamics, and annual census-based socioeconomic variables under a unified schema, and includes time-stamped annual aerial imagery. Building on HouseTS, we define standardized long-horizon forecasting tasks for univariate and multivariate prediction and benchmark 16 model families spanning statistical methods, classical machine learning, deep neural networks, and time-series foundation models in both zero-shot and fine-tuned modes. We also provide image-derived textual change annotations from multi-year aerial image sequences via a vision--language pipeline with LLM-as-judge and human verification to support scalable interpretability analyses. HouseTS is available on Kaggle, with code and documentation on GitHub.
Authors: Jiaxiang Chen, Zhuo Wang, Mingxi Zou, Qifan Wang, Zenglin Xu
Abstract: Retrieval-augmented generation (RAG) has been widely adopted to ground large language models (LLMs) in external knowledge, yet it remains largely underexplored for improving reasoning. Existing methods either rely on online exploration during inference or heuristic supervision over reasoning trajectories, but they fail to effectively accumulate and reuse past reasoning experience. We propose Guideline Forest, a retrieval-augmented reasoning framework that explicitly leverages experience to guide multi-step reasoning. The framework stores high-quality, label-consistent reasoning traces as reusable memory, retrieves relevant experiences for new problems, and induces them into structured guidelines that steer reasoning and enable controlled branching and aggregation. Experiments on mathematical (GSM8K, MATH-500) and programming (MBPP, HumanEval) benchmarks demonstrate consistent improvements over strong reasoning baselines, including CoT, ReAct, ToT, FoT, and AFlow. Further analyses show that experience retrieval, guideline-induced diversity, and stepwise aggregation are key to the framework's effectiveness. Beyond single-model reasoning, Guideline Forest generalizes to enhance diverse reasoning paradigms and supports multi-model collaboration, highlighting its flexibility and scalability.
Authors: Leheng Sheng, An Zhang, Zijian Wu, Weixiang Zhao, Changshuo Shen, Yi Zhang, Xiang Wang, Tat-Seng Chua
Abstract: Recent studies empirically reveal that large reasoning models (LRMs) can automatically allocate more reasoning strengths (i.e., the number of reasoning tokens) for harder problems, exhibiting difficulty-awareness for better task performance. While this automatic reasoning strength allocation phenomenon has been widely observed, its underlying mechanism remains largely unexplored. To this end, we provide explanations for this phenomenon from the perspective of model activations. We find evidence that LRMs pre-plan the reasoning strengths in their activations even before generation, with this reasoning strength causally controlled by the magnitude of a pre-allocated directional vector. Specifically, we show that the number of reasoning tokens is predictable solely based on the question activations using linear probes, indicating that LRMs estimate the required reasoning strength in advance. We then uncover that LRMs encode this reasoning strength through a pre-allocated directional vector embedded in the activations of the model, where the vector's magnitude modulates the reasoning strength. Subtracting this vector can lead to reduced reasoning token number and performance, while adding this vector can lead to increased reasoning token number and even improved performance. We further reveal that this direction vector consistently yields positive reasoning length prediction, and it modifies the logits of end-of-reasoning token to affect the reasoning length. Finally, we demonstrate two potential applications of our findings: overthinking behavior detection and enabling efficient reasoning on simple problems. Our work provides new insights into the internal mechanisms of reasoning in LRMs and offers practical tools for controlling their reasoning behaviors. Our code is available at https://github.com/AlphaLab-USTC/LRM-plans-CoT.
Authors: Ziqi Zhu, Tao Hu, Honglong Zhang, Dan Yang, Hangeng Chen, Mengran Zhang, Xilun Chen
Abstract: We present CID-GraphRAG (Conversational Intent-Driven Graph Retrieval-Augmented Generation), a novel framework that addresses the limitations of existing dialogue systems in maintaining both contextual coherence and goal-oriented progression in multi-turn customer service conversations. Unlike traditional RAG systems that rely solely on semantic similarity or static knowledge graphs, CID-GraphRAG constructs intent transition graphs from goal-achieved historical dialogues and implements a dual-retrieval mechanism that balances intent-based graph traversal with semantic search. This approach enables the system to simultaneously leverage both conversational intent flow patterns and contextual semantics, significantly improving retrieval quality and response quality. In extensive experiments on real-world customer service dialogues, we demonstrated that CID-GraphRAG significantly outperforms both semantic-based and intent-based baselines across automatic metrics, LLM-as-a-Judge evaluations and human evaluations, with relative gains of 11.4% in BLEU, 4.9% in ROUGE, and 5.9% in METEOR. Most notably, CID-GraphRAG achieves a 57.9% improvement in response quality according to LLM-as-a-Judge evaluations. These results demonstrate that integrating intent transition structures with semantic retrieval creates a synergistic effect that neither approach achieves independently, establishing CID-GraphRAG as an effective framework for real-world multi-turn dialogue systems in customer service and other knowledge-intensive domains.
Authors: Weizhi Tang, Kwabena Nuamah, Vaishak Belle
Abstract: While LLMs have demonstrated impressive abilities across various domains, they struggle with two major issues. The first is that LLMs trap themselves into local optima and the second is that they lack exhaustive coverage of the solution space. To investigate and improve these two issues, we propose Lyria, a neuro-symbolic reasoning framework building on the integration of LLMs, genetic algorithms, and symbolic systems, comprising 7 essential components. Through conducting extensive experiments with 4 LLMs across 3 types of problems, we demonstrated the efficacy of Lyria. Furthermore, with 7 additional ablation experiments, we further systematically analyzed and elucidated the factors that affect its performance. In addition, based on Lyria, we extend the ideas to the fine-tuning process of LLMs and introduce LAFT which enables a weaker model to imitate the reasoning process of a stronger model that reason under the Lyria reasoning framework. We demonstrate that the significant effectiveness of LAFT by conducting extensive experiments against 9 constructed baselines. We finally reveal the limitations and provide insights into future directions.
Authors: Bill Marino, Ari Juels
Abstract: There is growing interest in giving AI agents access to cryptocurrencies as well as to the smart contracts that transact them. But doing so, this position paper argues, could lead to formidable new vectors of AI harm. To support this argument, we first examine the unique properties of cryptocurrencies and smart contracts that could give rise to these new vectors of AI harm. Next, we describe each of these new vectors of AI harm in detail, providing a first-of-its-kind taxonomy. Finally, we conclude with a call for more technical research aimed at preventing and mitigating these new vectors of AI , thereby making it safer to endow AI agents with cryptocurrencies and smart contracts.
Authors: Rui Wang, Qihan Lin, Jiayu Liu, Qing Zong, Tianshi Zheng, Weiqi Wang, Yangqiu Song
Abstract: Prospect Theory (PT) models human decision-making under uncertainty, while epistemic markers (e.g., maybe) serve to express uncertainty in language. However, it remains largely unexplored whether Prospect Theory applies to contemporary Large Language Models and whether epistemic markers, which express human uncertainty, affect their decision-making behaviour. To address these research gaps, we design a three-stage experiment based on economic questionnaires. We propose a more general and precise evaluation framework to model LLMs' decision-making behaviour under PT, introducing uncertainty through the empirical probability values associated with commonly used epistemic markers in comparable contexts. We then incorporate epistemic markers into the evaluation framework based on their corresponding probability values to examine their influence on LLM decision-making behaviours. Our findings suggest that modelling LLMs' decision-making with PT is not consistently reliable, particularly when uncertainty is expressed in diverse linguistic forms. Our code is released in https://github.com/HKUST-KnowComp/MarPT.
Authors: Chentong Chen, Mengyuan Zhong, Ye Fan, Jialong Shi, Jianyong Sun
Abstract: LLM-based Automatic Heuristic Design (AHD) within Evolutionary Computation (EC) frameworks has shown promising results. However, its effectiveness is hindered by the use of static operators and the lack of knowledge accumulation mechanisms. We introduce HiFo-Prompt, a framework that guides LLMs with two synergistic prompting strategies: Foresight and Hindsight. Foresight-based prompts adaptively steer the search based on population dynamics, managing the exploration-exploitation trade-off. In addition, hindsight-based prompts mimic human expertise by distilling successful heuristics from past generations into fundamental, reusable design principles. This dual mechanism transforms transient discoveries into a persistent knowledge base, enabling the LLM to learn from its own experience. Empirical results demonstrate that HiFo-Prompt significantly outperforms state-of-the-art LLM-based AHD methods, generating higher-quality heuristics while achieving substantially faster convergence and superior query efficiency. Our code is available at https://github.com/Challenger-XJTU/HiFo-Prompt.
Authors: Wei Yang, Jinwei Xiao, Hongming Zhang, Qingyang Zhang, Yanna Wang, Bo Xu
Abstract: Recent advancements in Large Language Models (LLMs) have driven growing interest in LLM-based agents for complex planning tasks. To avoid costly agent training, many studies adopted memory mechanism that enhances LLM with offline experiences or online trajectory analysis. However, existing works focus on single-granularity memory derived from dynamic environmental interactions, which are inherently constrained by the quality of the collected experiences. This limitation, in turn, constrain the diversity of knowledge and the flexibility of planning. We propose Coarse-to-Fine Grounded Memory (\Ours{}), a novel framework that grounds coarse-to-fine memories with LLM, thereby fully leverage them for flexible adaptation to diverse scenarios. \Ours{} grounds environmental information into coarse-grained focus points to guide experience collection in training tasks, followed by grounding of actionable hybrid-grained tips from each experience. At inference, \Ours{} retrieves task-relevant experiences and tips to support planning. When facing environmental anomalies, the LLM grounds the current situation into fine-grained key information, enabling flexible self-QA reflection and plan correction.
Authors: Jo\~ao Paulo Nogueira, Wentao Sun, Alonso Silva, Laith Zumot
Abstract: Large reasoning language models are typically run with fixed inference budgets, which can waste computation or terminate reasoning prematurely. We introduce Certainty-Guided Reasoning (CGR), a model-agnostic adaptive inference procedure that periodically probes whether the current reasoning supports a confident final answer and terminates early once a target certainty threshold is reached, otherwise continuing until the end-of-thinking token or the budget limit. Certainty is estimated from the model's predicted probabilities over the answer tokens, yielding a lightweight stopping criterion. On AIME2025, CGR preserves baseline accuracy while reducing token usage, providing a tunable certainty-efficiency trade-off that can eliminate millions of tokens in aggregate. Across 64 random seeds, CGR exhibits consistent behavior. We also introduce a Grade metric that penalizes incorrect answers and permits abstention, capturing risk-sensitive performance. Results show that CGR improves Grade by abstaining when certainty remains low.
Authors: Tong Nie, Yuewen Mei, Yihong Tang, Junlin He, Jie Sun, Haotian Shi, Wei Ma, Jian Sun
Abstract: Adversarial scenario generation is a cost-effective approach for safety assessment of autonomous driving systems. However, existing methods are often constrained to a single, fixed trade-off between competing objectives such as adversariality and realism. This yields behavior-specific models that cannot be steered at inference time, lacking the efficiency and flexibility to generate tailored scenarios for diverse training and testing requirements. In view of this, we reframe the task of adversarial scenario generation as a multi-objective preference alignment problem and introduce a new framework named \textbf{S}teerable \textbf{A}dversarial scenario \textbf{GE}nerator (SAGE). SAGE enables fine-grained test-time control over the trade-off between adversariality and realism without any retraining. We first propose hierarchical group-based preference optimization, a data-efficient offline alignment method that learns to balance competing objectives by decoupling hard feasibility constraints from soft preferences. Instead of training a fixed model, SAGE fine-tunes two experts on opposing preferences and constructs a continuous spectrum of policies at inference time by linearly interpolating their weights. We provide theoretical justification for this framework through the lens of linear mode connectivity. Extensive experiments demonstrate that SAGE not only generates scenarios with a superior balance of adversariality and realism but also enables more effective closed-loop training of driving policies. Project page: https://tongnie.github.io/SAGE/.
Authors: Zijian Shao, Haiyang Shen, Mugeng Liu, Gecheng Fu, Yaoqi Guo, Yanfeng Wang, Yun Ma
Abstract: In clinical decision-making, predictive models face a persistent trade-off: accurate models are often opaque "black boxes," while interpretable methods frequently lack predictive precision or statistical grounding. In this paper, we challenge this dichotomy, positing that high predictive accuracy and high-quality descriptive explanations are not competing goals but synergistic outcomes of a deep, first-hand understanding of data. We propose the Reflective Cognitive Architecture (RCA), a novel framework designed to enable Large Language Models (LLMs) to learn directly from tabular data through experience and reflection. RCA integrates two core mechanisms: an iterative rules optimization process that refines logical argumentation by learning from prediction errors, and a distribution-aware rules check that grounds this logic in global statistical evidence to ensure robustness. We evaluated RCA against over 20 baselines - ranging from traditional machine learning to advanced reasoning LLMs and agents - across diverse medical datasets, including a proprietary real-world Catheter-Related Thrombosis (CRT) cohort. Crucially, to demonstrate real-world scalability, we extended our evaluation to two large-scale datasets. The results confirm that RCA achieves state-of-the-art predictive performance and superior robustness to data noise while simultaneously generating clear, logical, and evidence-based explanatory statements, maintaining its efficacy even at scale. The code is available at https://github.com/ssssszj/RCA.
Authors: Dongkyu Cho, Amy B. Z. Zhang, Bilel Fehri, Sheng Wang, Rumi Chunara, Hengrui Cai, Rui Song
Abstract: Chain-of-thought (CoT) reasoning exposes the intermediate thinking process of large language models (LLMs), yet verifying those traces at scale remains unsolved. In response, we introduce the idea of decision pivots-minimal, verifiable checkpoints that any correct reasoning path must visit. We hypothesize that correct reasoning, though stylistically diverse, converge on the same pivot set, while incorrect ones violate at least one pivot. Leveraging this property, we propose a self-training pipeline that (i) samples diverse reasoning paths and mines shared decision pivots, (ii) compresses each trace into pivot-focused short-path reasoning using an auxiliary verifier, and (iii) post-trains the model using its self-generated outputs. The proposed method aligns reasoning without ground truth reasoning data or external metrics. Experiments on standard benchmarks such as LogiQA, MedQA, and MATH500 show the effectiveness of our method.
Authors: Jiyuan Pei, Yi Mei, Jialin Liu, Mengjie Zhang, Xin Yao
Abstract: Recent neural solvers have demonstrated promising performance in learning to solve routing problems. However, existing studies are primarily based on one-off training on one or a set of predefined problem distributions and scales, i.e., tasks. When a new task arises, they typically rely on either zero-shot generalization, which may be poor due to the discrepancies between the new task and the training task(s), or fine-tuning the pretrained solver on the new task, which possibly leads to catastrophic forgetting of knowledge acquired from previous tasks. This paper explores a novel lifelong learning paradigm for neural VRP solvers, where multiple tasks with diverse distributions and scales arise sequentially over time. Solvers are required to effectively and efficiently learn to solve new tasks while maintaining their performance on previously learned tasks. Consequently, a novel framework called Lifelong Learning Router with Behavior Consolidation (LLR-BC) is proposed. LLR-BC consolidates prior knowledge effectively by aligning behaviors of the solver trained on a new task with the buffered ones in a decision-seeking way. To encourage more focus on crucial experiences, LLR-BC assigns greater consolidated weights to decisions with lower confidence. Extensive experiments on capacitated vehicle routing problems and traveling salesman problems demonstrate LLR-BC's effectiveness in training high-performance neural solvers in a lifelong learning setting, addressing the catastrophic forgetting issue, maintaining their plasticity, and improving zero-shot generalization ability.
Authors: Ziyang Zheng, Jiaying Zhu, Jingyi Zhou, Qiang Xu
Abstract: Learning to compute, the ability to model the functional behavior of a circuit graph, is a fundamental challenge for graph representation learning. Yet, the dominant paradigm is architecturally mismatched for this task. This flawed assumption, central to mainstream message passing neural networks (MPNNs) and their conventional Transformer-based counterparts, prevents models from capturing the position-aware, hierarchical nature of computation. To resolve this, we introduce TRACE, a new paradigm built on an architecturally sound backbone and a principled learning objective. First, TRACE employs a Hierarchical Transformer that mirrors the step-by-step flow of computation, providing a faithful architectural backbone that replaces the flawed permutation-invariant aggregation. Second, we introduce function shift learning, a novel objective that decouples the learning problem. Instead of predicting the complex global function directly, our model is trained to predict only the function shift, the discrepancy between the true global function and a simple local approximation that assumes input independence. We validate this paradigm on various circuits modalities, including Register Transfer Level graphs, And-Inverter Graphs and post-mapping netlists. Across a comprehensive suite of benchmarks, TRACE substantially outperforms all prior architectures. These results demonstrate that our architecturally-aligned backbone and decoupled learning objective form a more robust paradigm for the fundamental challenge of learning the functional behavior of a circuit graph.
Authors: Mingyi Luo, Ruichen Zhang, Xiangwang Hou, Jun Du, Chunxiao Jiang, Yong Ren, Dusit Niyato, Shiwen Mao
Abstract: The rapid advancement of large language models (LLMs) has enabled an emergence of agentic artificial intelligence (AI) with powerful reasoning and autonomous decision-making capabilities. This integration with edge computing has led to the development of Mobile Edge General Intelligence (MEGI), which brings real-time, privacy-preserving reasoning to the network edge. However, deploying LLM-based agentic AI reasoning in MEGI environments poses significant challenges due to the high computational demands of reasoning and the limited resources of edge devices. To address these challenges, we propose a joint optimization framework for efficient LLM reasoning deployment in MEGI. First, we systematically review enhancement methods to identify mechanisms suitable for edge adaptation. Subsequently, we present a distributed framework that synergizes reasoning enhancement via adaptive CoT prompting with scalable deployment through a distributed MoE architecture. An important innovation of this approach involves modeling reasoning depth as a dynamic network resource variable, which is optimized jointly with expert activation and transmission power. This mechanism allows the system to dynamically regulate expert networks and reasoning complexity according to task requirements and device capabilities. Experimental evaluations in mobile edge environments demonstrate that the proposed framework effectively balances reasoning quality and resource efficiency. The results show that with less than one second of additional inference time, both accuracy and latency satisfaction rate can reach 90\%, validating the practical viability of deploying sophisticated LLM reasoning in resource-constrained MEGI systems.
Authors: Bill Marino, Rosco Hunter, Christoph Schnabl, Zubair Jamali, Marinos Emmanouil Kalpakos, Mudra Kashyap, Isaiah Hinton, Alexa Hanson, Maahum Nazir, Felix Steffek, Hongkai Wen, Nicholas D. Lane
Abstract: As governments move to regulate AI, there is growing interest in using Large Language Models (LLMs) to assess whether or not an AI system complies with a given AI Regulation (AIR). However, there is presently no way to benchmark the performance of LLMs at this task. To fill this void, we introduce AIReg-Bench: the first open benchmark dataset designed to test how well LLMs can assess compliance with the EU AI Act (AIA). We created this dataset through a two-step process: (1) by prompting an LLM with carefully structured instructions, we generated 120 technical documentation excerpts (samples), each depicting a fictional, albeit plausible, AI system -- of the kind an AI provider might produce to demonstrate their compliance with AIR; (2) legal experts then reviewed and annotated each sample to indicate whether, and in what way, the AI system described therein violates specific Articles of the AIA. The resulting dataset, together with our evaluation of whether frontier LLMs can reproduce the experts' compliance labels, provides a starting point to understand the opportunities and limitations of LLM-based AIR compliance assessment tools and establishes a benchmark against which subsequent LLMs can be compared. The dataset and evaluation code are available at https://github.com/camlsys/aireg-bench.
Authors: Minh-Anh Nguye, Minh-Duc Nguyen, Ha Lan N. T., Kieu Hai Dang, Nguyen Tien Dong, Dung D. Le
Abstract: Large language models (LLMs) are increasingly adopted for automating survey paper generation \cite{wang2406autosurvey, liang2025surveyx, yan2025surveyforge,su2025benchmarking,wen2025interactivesurvey}. Existing approaches typically extract content from a large collection of related papers and prompt LLMs to summarize them directly. However, such methods often overlook the structural relationships among papers, resulting in generated surveys that lack a coherent taxonomy and a deeper contextual understanding of research progress. To address these shortcomings, we propose \textbf{SurveyG}, an LLM-based agent framework that integrates \textit{hierarchical citation graph}, where nodes denote research papers and edges capture both citation dependencies and semantic relatedness between their contents, thereby embedding structural and contextual knowledge into the survey generation process. The graph is organized into three layers: \textbf{Foundation}, \textbf{Development}, and \textbf{Frontier}, to capture the evolution of research from seminal works to incremental advances and emerging directions. By combining horizontal search within layers and vertical depth traversal across layers, the agent produces multi-level summaries, which are consolidated into a structured survey outline. A multi-agent validation stage then ensures consistency, coverage, and factual accuracy in generating the final survey. Experiments, including evaluations by human experts and LLM-as-a-judge, demonstrate that SurveyG outperforms state-of-the-art frameworks, producing surveys that are more comprehensive and better structured to the underlying knowledge taxonomy of a field.
Authors: Zixuan Qin, Qingchen Yu, Kunlin Lyu, Zhaoxin Fan, Yifan Sun
Abstract: Large Language Models (LLMs) have become foundational tools in natural language processing, powering a wide range of applications and research. Many studies have shown that LLMs share significant similarities with the human brain. Recent neuroscience research has found that a small subset of biological neurons in the human brain are crucial for core cognitive functions, which raises a fundamental question: do LLMs also contain a small subset of critical neurons? In this paper, we investigate this question by proposing a Perturbation-based Causal Identification of Critical Neurons method to systematically locate such critical neurons in LLMs. Our findings reveal three key insights: (1) LLMs contain ultra-sparse critical neuron sets. Disrupting these critical neurons can cause a 72B-parameter model with over 1.1 billion neurons to completely collapse, with perplexity increasing by up to 20 orders of magnitude; (2) These critical neurons are not uniformly distributed, but tend to concentrate in the outer layers, particularly within the MLP down\_proj components; (3) Performance degradation exhibits sharp phase transitions, rather than a gradual decline, when these critical neurons are disrupted. Through comprehensive experiments across diverse model architectures and scales, we provide deeper analysis of these phenomena and their implications for LLM robustness and interpretability. These findings can offer guidance for developing more robust model architectures and improving deployment security in safety-critical applications. Our code is available at https://github.com/qqqqqqqzx/The-Achilles-Heel-of-LLMs.
URLs: https://github.com/qqqqqqqzx/The-Achilles-Heel-of-LLMs.
Authors: Yangqin Jiang, Chao Huang
Abstract: With the advancement of multimodal large language models (MLLMs), building GUI agent systems has become an increasingly promising direction--especially for mobile platforms, given their rich app ecosystems and intuitive touch interactions. Yet mobile GUI agents face a critical dilemma: truly on-device models (4B or smaller) lack sufficient performance, while capable models (starting from 7B) are either too large for mobile deployment or prohibitively costly (e.g., cloud-only closed-source MLLMs). To resolve this, we propose OpenPhone, a mobile GUI agent system that leverages device-cloud collaboration to tap the cost-efficiency of on device models and the high capability of cloud models, while avoiding their drawbacks. Specifically, OpenPhone enhances Qwen2.5-VL-3B via two-stage SFT->GRPO training on synthetic GUI data for strong decision-making, integrates an efficient long-reasoning and memory management mechanism to utilize historical interactions under tight resources, and defaults to on-device execution--only escalating challenging subtasks to the cloud via real-time complexity assessment. Experiments on the online AndroidLab benchmark and diverse apps show OpenPhone matches or nears larger models, with a significant reduction in cloud costs.
Authors: Chenrui Shi, Zedong Yu, Zhi Gao, Ruining Feng, Enqi Liu, Yuwei Wu, Yunde Jia, Liuyu Xiang, Zhaofeng He, Qing Li
Abstract: Vision language models (VLMs) have advanced graphical user interface (GUI) task automation but still lag behind humans. We hypothesize this gap stems from missing core GUI knowledge, which existing training schemes (such as supervised fine tuning and reinforcement learning) alone cannot fully address. By analyzing common failure patterns in GUI task execution, we distill GUI knowledge into three dimensions: (1) interface knowledge about widget functions, layout semantics, and system states; (2) interaction knowledge about GUI interaction types and effects; and (3) procedure knowledge of task objectives and workflow sequences. We further introduce GUI Knowledge Bench, a benchmark with multiple-choice and yes/no questions across six platforms (Web, Android, MacOS, Windows, Linux, IOS) and 292 applications. Our evaluation indicates that current VLMs are generally aware of the functions of individual widgets, but lack the GUI-specific knowledge required to track system states, adhere to GUI interaction conventions, and assess task completion progress. Experiments on real-world GUI tasks further validate the close link between GUI knowledge and task success. By providing a structured framework for assessing GUI knowledge, our work supports the selection of VLMs with greater potential prior to downstream training and provides insights for building more capable GUI agents.
Authors: Qiusi Zhan, Hyeonjeong Ha, Rui Yang, Sirui Xu, Hanyang Chen, Liang-Yan Gui, Yu-Xiong Wang, Huan Zhang, Heng Ji, Daniel Kang
Abstract: Recent advances in Vision-Language Models (VLMs) have propelled embodied agents by enabling direct perception, reasoning, and planning task-oriented actions from visual inputs. However, such vision-driven embodied agents open a new attack surface: visual backdoor attacks, where the agent behaves normally until a visual trigger appears in the scene, then persistently executes an attacker-specified multi-step policy. We introduce BEAT, the first framework to inject such visual backdoors into VLM-based embodied agents using objects in the environments as triggers. Unlike textual triggers, object triggers exhibit wide variation across viewpoints and lighting, making them difficult to implant reliably. BEAT addresses this challenge by (1) constructing a training set that spans diverse scenes, tasks, and trigger placements to expose agents to trigger variability, and (2) introducing a two-stage training scheme that first applies supervised fine-tuning (SFT) and then our novel Contrastive Trigger Learning (CTL). CTL formulates trigger discrimination as preference learning between trigger-present and trigger-free inputs, explicitly sharpening the decision boundaries to ensure precise backdoor activation. Across various embodied agent benchmarks and VLMs, BEAT achieves attack success rates up to 80%, while maintaining strong benign task performance, and generalizes reliably to out-of-distribution trigger placements. Notably, compared to naive SFT, CTL boosts backdoor activation accuracy up to 39% under limited backdoor data. These findings expose a critical yet unexplored security risk in VLM-based embodied agents, underscoring the need for robust defenses before real-world deployment.
Authors: Tom\'a\v{s} \v{C}\'i\v{z}ek, Martin Balko, Martin Schmid
Abstract: Proof-Number Search is a best-first search algorithm with many successful applications, especially in game solving. As large-scale computing clusters become increasingly accessible, parallelization is a natural way to accelerate computation. However, existing parallel versions of Proof-Number Search are known to scale poorly on many CPU cores. Using two parallelized levels and shared information among workers, we present the first massively parallel version of Proof-Number Search that scales efficiently even on a large number of CPUs. We apply our solver, enhanced with Grundy numbers for reducing game trees of impartial games, to the Sprouts game, a case study motivated by the long-standing Sprouts Conjecture. Our algorithm achieves 332.9$\times$ speedup on 1024 cores, significantly improving previous parallelizations and outperforming the state-of-the-art Sprouts solver GLOP by four orders of magnitude in runtime while generating proofs 1,000$\times$ more complex. Despite exponential growth in game tree size, our solver verified the Sprouts Conjecture for 42 new positions, nearly doubling the number of known outcomes.
Authors: Julia Santaniello, Matthew Russell, Benson Jiang, Donatello Sassaroli, Robert Jacob, Jivko Sinapov
Abstract: Reinforcement Learning from Human Feedback (RLHF) is a methodology that aligns agent behavior with human preferences by integrating user feedback into the agent's training process. This paper introduces a framework that guides agent training through implicit neural signals, with a focus on the neural classification problem. Our work presents and releases a novel dataset of functional near-infrared spectroscopy (fNIRS) recordings collected from 25 human participants across three domains: Pick-and-Place Robot, Lunar Lander, and Flappy Bird. We train multiple classifiers to predict varying levels of agent performance (optimal, suboptimal, or worst-case) from windows of preprocessed fNIRS features, achieving an average F1 score of 67% for binary and 46% for multi-class classification across conditions and domains. We also train multiple regressors to predict the degree of deviation between an agent's chosen action and a set of near-optimal policy actions, providing a continuous measure of performance. Finally, we evaluate cross-subject generalization and show that fine-tuning pre-trained models with a small sample of subject-specific data increases average F1 scores by 17% and 41% for binary and multi-class models, respectively. Our results demonstrate that mapping implicit fNIRS signals to agent performance is feasible and can be improved, laying the foundation for future Reinforcement Learning from Neural Feedback (RLNF) systems.
Authors: Chandrachur Bhattacharya, Sibendu Som
Abstract: AI Scientific Assistant Core (AISAC) is a transparent, modular multi-agent runtime developed at Argonne National Laboratory to support long-horizon, evidence-grounded scientific reasoning. Rather than proposing new agent algorithms or claiming autonomous scientific discovery, AISAC contributes a governed execution substrate that operationalizes key requirements for deploying agentic AI in scientific practice, including explicit role semantics, budgeted context management, traceable execution, and reproducible interaction with tools and knowledge. AISAC enforces four structural guarantees for scientific reasoning: (1) declarative agent registration with runtime-enforced role semantics and automatic system prompt generation; (2) budgeted orchestration via explicit per-turn context and delegation depth limits; (3) role-aligned memory access across episodic, dialogue, and evidence layers; and (4) trace-driven transparency through persistent execution records and a live event-stream interface. These guarantees are implemented through hybrid persistent memory (SQLite and dual FAISS indices), governed retrieval with agent-scoped RAG, structured tool execution with schema validation, and a configuration-driven bootstrap mechanism that enables project specific extension without modifying the shared core. AISAC is currently deployed across multiple scientific workflows at Argonne, including combustion science, materials research, and energy process safety, demonstrating its use as a reusable substrate for domain-specialized AI scientific assistants.
Authors: Shaoyin Ma, Chenggong Hu, Huiqiong Wang, Li Sun, Mingli Song, Jie Song
Abstract: Building effective LLM agents increasingly requires selecting appropriate AI models as tools from large open repositories (e.g., HuggingFace with > 2M models) based on natural language requests. Unlike invoking a fixed set of API tools, repository-scale model selection must handle massive, evolving candidates with incomplete metadata. Existing approaches incorporate full model descriptions into prompts, resulting in prompt bloat, excessive token costs, and limited scalability. To address these issues, we propose HuggingR$^4$, the first framework to recast model selection as an iterative reasoning process rather than one-shot retrieval. By synergistically integrating Reasoning, Retrieval, Refinement, and Reflection, HuggingR$^4$ progressively decomposes user intent, retrieves candidates through multi-round deliberation, refines selections via fine-grained analysis, and validates results through reflection. To facilitate rigorous evaluation, we introduce a large-scale benchmark comprising 14,399 diverse user requests across 37 task categories. Experiments demonstrate that HuggingR$^4$ achieves 92.03% workability and 82.46% reasonability-outperforming current state-of-the-art baselines by 26.51% and 33.25%, respectively, while reducing token consumption by $6.9 \times$.
Authors: Changxin Huang, Lv Tang, Zhaohuan Zhan, Lisha Yu, Runhao Zeng, Zun Liu, Zhengjie Wang, Jianqiang Li
Abstract: Vision-and-Language Navigation (VLN) requires agents to autonomously navigate complex environments via visual images and natural language instructions--remains highly challenging. Recent research on enhancing language-guided navigation reasoning using pre-trained large language models (LLMs) has shown promising prospects. However, the reasoning of such methods is limited to the linguistic modality, lacking visual reasoning capabilities. Moreover, existing reasoning modules are optimized separately from navigation policies, leading to incompatibility and potential conflicts in optimization objectives.To tackle these challenges, we introduce UNeMo, a novel framework designed for the collaborative optimization of visual state reasoning and navigational decision-making. It introduces a Multimodal World Model (MWM) that takes visual features, language instructions, and navigational actions as inputs to jointly predict subsequent visual states, enabling cross-modal reasoning. Via a Hierarchical Prediction-Feedback (HPN) mechanism, MWM collaborates with navigation policies: the first layer generates actions using current vision-and-language features; MWM then infers post-action visual states to guide the second layer's fine-grained decisions. This forms a dynamic bidirectional promotion mechanism where MWM reasoning optimizes navigation policies, while policy decisions feedback to improve MWM's reasoning accuracy. Experiments on R2R and REVERIE datasets show UNeMo outperforms state-of-the-art methods by 2.1% and 0.7% in navigation accuracy for unseen scenes, validating its effectiveness.
Authors: Haebin Seong, Sungmin Kim, Yongjun Cho, Myunchul Joe, Geunwoo Kim, Yubeen Park, Sunhoo Kim, Yoonshik Kim, Suhwan Choi, Jaeyoon Jung, Jiyong Youn, Jinmyung Kwak, Sunghee Ahn, Jaemin Lee, Younggil Do, Seungyeop Yi, Woojin Cheong, Minhyeok Oh, Minchan Kim, Yoonseok Kang, Seongjae Kang, Samwoo Seong, Youngjae Yu, Yunsung Lee
Abstract: While current navigation benchmarks prioritize task success in simplified settings, they neglect the multidimensional economic constraints essential for the real-world commercialization of autonomous delivery systems. We introduce CostNav, an Economic Navigation Benchmark that evaluates physical AI agents through comprehensive economic cost-revenue analysis aligned with real-world business operations. By integrating industry-standard data - such as SEC filings and AIS injury reports - with Isaac Sim's detailed collision and cargo dynamics, CostNav transcends simple task completion to accurately evaluate business value in complex, real-world scenarios. To our knowledge, CostNav is the first work to quantitatively expose the gap between navigation research metrics and commercial viability, revealing that optimizing for task success on a simplified task fundamentally differs from optimizing for real-world economic deployment. Our evaluation of rule-based Nav2 navigation shows that current approaches are not economically viable: the contribution margin is -22.81/run (AMCL) and -12.87/run (GPS), resulting in no break-even point. We challenge the community to develop navigation policies that achieve economic viability on CostNav. We remain method-agnostic, evaluating success solely on the metric of cost rather than the underlying architecture. All resources are available at https://github.com/worv-ai/CostNav.
Authors: Kevin Lee, Russell Spiewak, James Walsh
Abstract: Scientific reasoning through Large Language Models in heliophysics involves more than just recalling facts: it requires incorporating physical assumptions, maintaining consistent units, and providing clear scientific formats through coordinated approaches. To address these challenges, we present Reasoning With a Star, a newly contributed heliophysics dataset applicable to reasoning; we also provide an initial benchmarking approach. Our data are constructed from National Aeronautics and Space Administration & University Corporation for Atmospheric Research Living With a Star summer school problem sets and compiled into a readily consumable question-and-answer structure with question contexts, reasoning steps, expected answer type, ground-truth targets, format hints, and metadata. A programmatic grader checks the predictions using unit-aware numerical tolerance, symbolic equivalence, and schema validation. We benchmark a single-shot baseline and four multi-agent patterns, finding that decomposing workflows through systems engineering principles outperforms direct prompting on problems requiring deductive reasoning rather than pure inductive recall.
Authors: Simon S\"uwer, Kester Bagemihl, Sylvie Baier, Lucia Dicunta, Markus List, Jan Baumbach, Andreas Maier, Fernando M. Delgado-Chaves
Abstract: Repurposing approved drugs offers a time-efficient and cost-effective alternative to traditional drug development. However, in silico prediction of repurposing candidates is challenging and requires the effective collaboration of specialists in various fields, including pharmacology, medicine, biology, and bioinformatics. Fragmented, specialized algorithms and tools often address only narrow aspects of the overall problem. Heterogeneous, unstructured data landscapes require the expertise of specialized users. Hence, these data services do not integrate smoothly across workflows. With ChatDRex, we present a conversation-based, multi-agent system that facilitates the execution of complex bioinformatic analyses aiming for network-based drug repurposing prediction. It builds on the integrated systems medicine knowledge graph (NeDRex KG). ChatDRex provides natural language access to its extensive biomedical knowledge base. It integrates bioinformatics agents for network analysis, literature mining, and drug repurposing. These are complemented by agents that evaluate functional coherence for in silico validation. Its flexible multi-agent design assigns specific tasks to specialized agents, including query routing, data retrieval, algorithm execution, and result visualization. A dedicated reasoning module keeps the user in the loop and allows for hallucination detection. By enabling physicians and researchers without computer science expertise to control complex analyses with natural language, ChatDRex democratizes access to bioinformatics as an important resource for drug repurposing. It enables clinical experts to generate hypotheses and explore drug repurposing opportunities, ultimately accelerating the discovery of novel therapies and advancing personalized medicine and translational research. ChatDRex is publicly available at apps.cosy.bio/chatdrex.
Authors: Matt Wilson
Abstract: We establish a correspondence between equivalence classes of agent-state policies for deterministic POMDPs and one-input process functions (the classical-deterministic limit of higher-order quantum operations). We use this correspondence to build a bridge between the agent-environment interaction in artificial intelligence, causal structure in the foundations of physics, and logic in computer science. We construct a *-autonomous category PF of types which supports an interpretation of one-step evaluation of policies, and multi-agent observation constraints, into cuts and monoidal products. In terms of types, we develop the correspondence further by identifying observation-independent decentralised POMDPs as the natural domain for the multi-input process functions used to model indefinite causality. We then prove a strict separation between general multi-input process function and definite-ordered process function performance on such dec-POMDPs, by finding an instance for which policies utilizing an indefinite causal structure can achieve greater finite-horizon rewards than policies which are restricted to a fixed background causal structure.
Authors: Anselm Paulus, Ilia Kulikov, Brandon Amos, R\'emi Munos, Ivan Evtimov, Kamalika Chaudhuri, Arman Zharmagambetov
Abstract: Ensuring the safety of language models (LMs) while maintaining their usefulness remains a critical challenge in AI alignment. Current approaches rely on sequential adversarial training: generating adversarial prompts and fine-tuning LMs to defend against them. We introduce a different paradigm: framing safety alignment as a non-zero-sum game between an Attacker LM and a Defender LM trained jointly via online reinforcement learning. Each LM continuously adapts to the other's evolving strategies, driving iterative improvement. Our method uses a preference-based reward signal derived from pairwise comparisons instead of point-wise scores, providing more robust supervision and potentially reducing reward hacking. Our RL recipe, AdvGame, shifts the Pareto frontier of safety and utility, yielding a Defender LM that is simultaneously more helpful and more resilient to adversarial attacks. In addition, the resulting Attacker LM converges into a strong, general-purpose red-teaming agent that can be directly deployed to probe arbitrary target models.
Authors: Yu Li, Tian Lan, Zhengling Qi
Abstract: Direct Preference Optimization (DPO) and its variants have become standard for aligning Large Language Models due to their simplicity and offline stability. However, we identify two fundamental limitations. First, the optimal policy depends on arbitrary modeling choices (scalarization function, reference policy), yielding behavior reflecting parameterization artifacts rather than true preferences. Second, treating response generation in isolation fails to leverage comparative information in pairwise data, leaving the model's capacity for intrinsic self-reflection untapped. To address it, we propose Intrinsic Self-reflective Preference Optimization (InSPO), deriving a globally optimal policy conditioning on both context and alternative responses. We prove this formulation superior to DPO/RLHF while guaranteeing invariance to scalarization and reference choices. InSPO serves as a plug-and-play enhancement without architectural changes or inference overhead. Experiments demonstrate consistent improvements in win rates and length-controlled metrics, validating that unlocking self-reflection yields more robust, human-aligned LLMs. Our Code is available at https://github.com/Skylanding/InSPO.
Authors: Nuoya Xiong, Yuhang Zhou, Hanqing Zeng, Zhaorun Chen, Furong Huang, Shuchao Bi, Lizhu Zhang, Zhuokai Zhao
Abstract: Large language models (LLMs) exhibit strengths across diverse domains. However, achieving strong performance across these domains with a single general-purpose model typically requires scaling to sizes that are prohibitively expensive to train and deploy. On the other hand, while smaller domain-specialized models are much more efficient, they struggle to generalize beyond their training distributions. To address this dilemma, we propose FusionRoute, a robust and effective token-level multi-LLM collaboration framework in which a lightweight router simultaneously (i) selects the most suitable expert at each decoding step and (ii) contributes a complementary logit that refines or corrects the selected expert's next-token distribution via logit addition. Unlike existing token-level collaboration methods that rely solely on fixed expert outputs, we provide a theoretical analysis showing that pure expert-only routing is fundamentally limited: unless strong global coverage assumptions hold, it cannot in general realize the optimal decoding policy. By augmenting expert selection with a trainable complementary generator, FusionRoute expands the effective policy class and enables recovery of optimal value functions under mild conditions. Empirically, across both Llama-3 and Gemma-2 families and diverse benchmarks spanning mathematical reasoning, code generation, and instruction following, FusionRoute outperforms both sequence- and token-level collaboration, model merging, and direct fine-tuning, while remaining competitive with domain experts on their respective tasks.
Authors: Kabir Swain, Sijie Han, Ayush Raina, Jin Zhang, Shuang Li, Michael Stopa, Antonio Torralba
Abstract: As large language models (LLMs) continue to improve in reasoning and decision-making, there is a growing need for realistic and interactive environments where their abilities can be rigorously evaluated. We present VirtualEnv, a next-generation simulation platform built on Unreal Engine 5 that enables fine-grained benchmarking of LLMs in embodied and interactive scenarios. VirtualEnv supports rich agent-environment interactions, including object manipulation, navigation, and adaptive multi-agent collaboration, as well as game-inspired mechanics like escape rooms and procedurally generated environments. We provide a user-friendly API built on top of Unreal Engine, allowing researchers to deploy and control LLM-driven agents using natural language instructions. We integrate large-scale LLMs and vision-language models (VLMs), such as GPT-based models, to generate novel environments and structured tasks from multimodal inputs. Our experiments benchmark the performance of several popular LLMs across tasks of increasing complexity, analyzing differences in adaptability, planning, and multi-agent coordination. We also describe our methodology for procedural task generation, task validation, and real-time environment control. VirtualEnv is released as an open-source platform, we aim to advance research at the intersection of AI and gaming, enable standardized evaluation of LLMs in embodied AI settings, and pave the way for future developments in immersive simulations and interactive entertainment.
Authors: Hongyu Lin, Samer Abdallah, Makar Valentinov, Paul Brennan, Elijah Kagan, Christoph M. Wintersteiger, Denis Ignatovich, Grant Passmore
Abstract: Large Language Models (LLMs) have shown strong performance on code understanding tasks, yet they fundamentally lack the ability to perform precise, exhaustive mathematical reasoning about program behavior. Existing benchmarks either focus on mathematical proof automation, largely disconnected from real-world software, or on engineering tasks that do not require semantic rigor. We present CodeLogician, a neurosymbolic agent for precise analysis of software logic, integrated with ImandraX, an industrial automated reasoning engine deployed in financial markets and safety-critical systems. Unlike prior approaches that use formal methods primarily to validate LLM outputs, CodeLogician uses LLMs to construct explicit formal models of software systems, enabling automated reasoning to answer rich semantic questions beyond binary verification outcomes. To rigorously evaluate mathematical reasoning about software logic, we introduce code-logic-bench, a benchmark targeting the middle ground between theorem proving and software engineering benchmarks. It measures reasoning correctness about program state spaces, control flow, coverage constraints, and edge cases, with ground truth defined via formal modeling and region decomposition. Comparing LLM-only reasoning against LLMs augmented with CodeLogician, formal augmentation yields substantial improvements, closing a 41-47 percentage point gap in reasoning accuracy. These results demonstrate that neurosymbolic integration is essential for scaling program analysis toward rigorous, autonomous software understanding.
Authors: Xianzhe Meng, Qiangsheng Zeng, Ling Luo, Qinghan Yang, Jiarui Hao, Wenbo Wu, Qinyu Wang, Rui Yin, Lin Qi, Renzhi Lu
Abstract: Training stability is typically regarded as a prerequisite for reliable optimization in large language models. In this work, we analyze how stabilizing training dynamics affects the induced generation distribution. We show that under standard maximum likelihood training, stable parameter trajectories lead stationary solutions to approximately minimize the forward KL divergence to the empirical distribution, while implicitly reducing generative entropy. As a consequence, the learned model can concentrate probability mass on a limited subset of empirical modes, exhibiting systematic degeneration despite smooth loss convergence. We empirically validate this effect using a controlled feedback-based training framework that stabilizes internal generation statistics, observing consistent low-entropy outputs and repetitive behavior across architectures and random seeds. It indicates that optimization stability and generative expressivity are not inherently aligned, and that stability alone is an insufficient indicator of generative quality.
Authors: Yongxin Deng, Zhen Fang, Sharon Li, Ling Chen
Abstract: Hallucination detection is critical for deploying large language models (LLMs) in real-world applications. Existing hallucination detection methods achieve strong performance when the training and test data come from the same domain, but they suffer from poor cross-domain generalization. In this paper, we study an important yet overlooked problem, termed generalizable hallucination detection (GHD), which aims to train hallucination detectors on data from a single domain while ensuring robust performance across diverse related domains. In studying GHD, we simulate multi-turn dialogues following LLMs' initial response and observe an interesting phenomenon: hallucination-initiated multi-turn dialogues universally exhibit larger uncertainty fluctuations than factual ones across different domains. Based on the phenomenon, we propose a new score SpikeScore, which quantifies abrupt fluctuations in multi-turn dialogues. Through both theoretical analysis and empirical validation, we demonstrate that SpikeScore achieves strong cross-domain separability between hallucinated and non-hallucinated responses. Experiments across multiple LLMs and benchmarks demonstrate that the SpikeScore-based detection method outperforms representative baselines in cross-domain generalization and surpasses advanced generalization-oriented methods, verifying the effectiveness of our method in cross-domain hallucination detection.
Authors: Jarrod Barnes
Abstract: As large language models (LLMs) improve, so do their offensive applications: frontier agents now generate working exploits for under $50 in compute (Heelan, 2026). Defensive incident response (IR) agents must keep pace, but existing benchmarks conflate action execution with correct execution, hiding calibration failures when agents process adversarial evidence. We introduce OpenSec, a dual-control reinforcement learning (RL) environment that evaluates IR agents under realistic prompt injection scenarios with execution-based scoring: time-to-first-containment (TTFC), evidence-gated action rate (EGAR), blast radius, and per-tier injection violation rates. Evaluating four frontier models on 40 standard-tier episodes each, we find consistent over-triggering: GPT-5.2 executes containment in 100% of episodes with 82.5% false positive rate, acting at step 4 before gathering sufficient evidence. Claude Sonnet 4.5 shows partial calibration (62.5% containment, 45% FP, TTFC of 10.6), suggesting calibration is not reliably present across frontier models. All models correctly identify the ground-truth threat when they act; the calibration gap is not in detection but in restraint. Code available at https://github.com/jbarnes850/opensec-env.
Authors: Jacek Duszenko
Abstract: Reasoning models frequently agree with incorrect user suggestions -- a behavior known as sycophancy. However, it is unclear where in the reasoning trace this agreement originates and how strong the commitment is. We introduce \emph{sycophantic anchors} -- sentences identified via counterfactual analysis that commit models to user agreement. Across four reasoning models spanning three architecture families (Llama, Qwen, Falcon-hybrid) and 1.5B--8B parameters, we analyze over 200,000 counterfactual rollouts and show that linear probes reliably detect sycophantic anchors (74--85\% balanced accuracy), outperforming text-only baselines at high commitment levels -- confirming they capture internal states beyond surface vocabulary. Regressors further predict commitment strength from activations ($R^2$ up to 0.74). We observe a consistent asymmetry: sycophancy leaves a stronger mechanistic footprint than correct reasoning. We also find that sycophancy builds gradually during generation rather than being determined by the prompt. These findings enable sentence-level detection and quantification of model misalignment mid-inference.
Authors: Chentong Chen, Mengyuan Zhong, Ye Fan, Jialong Shi, Jianyong Sun
Abstract: Although Large Language Models have advanced Automated Heuristic Design, treating algorithm evolution as a monolithic text generation task overlooks the coupling between discrete algorithmic structures and continuous numerical parameters. Consequently, existing methods often discard promising algorithms due to uncalibrated constants and suffer from premature convergence resulting from simple similarity metrics. To address these limitations, we propose TIDE, a Tuning-Integrated Dynamic Evolution framework designed to decouple structural reasoning from parameter optimization. TIDE features a nested architecture where an outer parallel island model utilizes Tree Similarity Edit Distance to drive structural diversity, while an inner loop integrates LLM-based logic generation with a differential mutation operator for parameter tuning. Additionally, a UCB-based scheduler dynamically prioritizes high-yield prompt strategies to optimize resource allocation. Extensive experiments across nine combinatorial optimization problems demonstrate that TIDE discovers heuristics that significantly outperform state-of-the-art baselines in solution quality while achieving improved search efficiency and reduced computational costs.
Authors: Ruian Tie, Wenbo Xiong, Zhengyu Shi, Xinyu Su, Chenyu jiang, Libo Wu, Hao Li
Abstract: Conventional supervised climate downscaling struggles to generalize to Global Climate Models (GCMs) due to the lack of paired training data and inherent domain gaps relative to reanalysis. Meanwhile, current zero-shot methods suffer from physical inconsistencies and vanishing gradient issues under large scaling factors. We propose Zero-Shot Statistical Downscaling (ZSSD), a zero-shot framework that performs statistical downscaling without paired data during training. ZSSD leverages a Physics-Consistent Climate Prior learned from reanalysis data, conditioned on geophysical boundaries and temporal information to enforce physical validity. Furthermore, to enable robust inference across varying GCMs, we introduce Unified Coordinate Guidance. This strategy addresses the vanishing gradient problem in vanilla DPS and ensures consistency with large-scale fields. Results show that ZSSD significantly outperforms existing zero-shot baselines in 99th percentile errors and successfully reconstructs complex weather events, such as tropical cyclones, across heterogeneous GCMs.
Authors: Mingqian Feng, Xiaodong Liu, Weiwei Yang, Chenliang Xu, Christopher White, Jianfeng Gao
Abstract: Large Language Models (LLMs) are typically evaluated for safety under single-shot or low-budget adversarial prompting, which underestimates real-world risk. In practice, attackers can exploit large-scale parallel sampling to repeatedly probe a model until a harmful response is produced. While recent work shows that attack success increases with repeated sampling, principled methods for predicting large-scale adversarial risk remain limited. We propose a scaling-aware Best-of-N estimation of risk, SABER, for modeling jailbreak vulnerability under Best-of-N sampling. We model sample-level success probabilities using a Beta distribution, the conjugate prior of the Bernoulli distribution, and derive an analytic scaling law that enables reliable extrapolation of large-N attack success rates from small-budget measurements. Using only n=100 samples, our anchored estimator predicts ASR@1000 with a mean absolute error of 1.66, compared to 12.04 for the baseline, which is an 86.2% reduction in estimation error. Our results reveal heterogeneous risk scaling profiles and show that models appearing robust under standard evaluation can experience rapid nonlinear risk amplification under parallel adversarial pressure. This work provides a low-cost, scalable methodology for realistic LLM safety assessment. We will release our code and evaluation scripts upon publication to future research.
Authors: Xinyi Ke, Kai Li, Junliang Xing, Yifan Zhang, Jian Cheng
Abstract: Large language models (LLMs) have enabled rapid progress in automatic heuristic discovery (AHD), yet most existing methods are predominantly limited by static evaluation against fixed instance distributions, leading to potential overfitting and poor generalization under distributional shifts. We propose Algorithm Space Response Oracles (ASRO), a game-theoretic framework that reframes heuristic discovery as a program level co-evolution between solver and instance generator. ASRO models their interaction as a two-player zero-sum game, maintains growing strategy pools on both sides, and iteratively expands them via LLM-based best-response oracles against mixed opponent meta-strategies, thereby replacing static evaluation with an adaptive, self-generated curriculum. Across multiple combinatorial optimization domains, ASRO consistently outperforms static-training AHD baselines built on the same program search mechanisms, achieving substantially improved generalization and robustness on diverse and out-of-distribution instances.
Authors: Senkang Hu, Yong Dai, Yuzhi Zhao, Yihang Tao, Yu Guo, Zhengru Fang, Sam Tak Wu Kwong, Yuguang Fang
Abstract: Agentic reasoning enables large reasoning models (LRMs) to dynamically acquire external knowledge, but yet optimizing the retrieval process remains challenging due to the lack of dense, principled reward signals. In this paper, we introduce InfoReasoner, a unified framework that incentivizes effective information seeking via a synthetic semantic information gain reward. Theoretically, we redefine information gain as uncertainty reduction over the model's belief states, establishing guarantees, including non-negativity, telescoping additivity, and channel monotonicity. Practically, to enable scalable optimization without manual retrieval annotations, we propose an output-aware intrinsic estimator that computes information gain directly from the model's output distributions using semantic clustering via bidirectional textual entailment. This intrinsic reward guides the policy to maximize epistemic progress, enabling efficient training via Group Relative Policy Optimization (GRPO). Experiments across seven question-answering benchmarks demonstrate that InfoReasoner consistently outperforms strong retrieval-augmented baselines, achieving up to 5.4% average accuracy improvement. Our work provides a theoretically grounded and scalable path toward agentic reasoning with retrieval. The code is available at https://github.com/dl-m9/InfoReasoner
Authors: Nikita Benkovich, Vitalii Valkov
Abstract: Large language models have demonstrated strong capabilities in individual software engineering tasks, yet most autonomous systems still treat issue resolution as a monolithic or pipeline-based process. In contrast, real-world software development is organized as a collaborative activity carried out by teams following shared methodologies, with clear role separation, communication, and review. In this work, we present a fully automated multi-agent system that explicitly models software engineering as an organizational process, replicating the structure of an engineering team. Built on top of agyn, an open-source platform for configuring agent teams, our system assigns specialized agents to roles such as coordination, research, implementation, and review, provides them with isolated sandboxes for experimentation, and enables structured communication. The system follows a defined development methodology for working on issues, including analysis, task specification, pull request creation, and iterative review, and operates without any human intervention. Importantly, the system was designed for real production use and was not tuned for SWE-bench. When evaluated post hoc on SWE-bench 500, it resolves 72.2% of tasks, outperforming single-agent baselines using comparable language models. Our results suggest that replicating team structure, methodology, and communication is a powerful paradigm for autonomous software engineering, and that future progress may depend as much on organizational design and agent infrastructure as on model improvements.
Authors: Pengrui Lu, Shiqi Zhang, Yunzhong Hou, Lyumanshan Ye, Chaoyi Huang, Zixi Chen, Ji Zeng, Hantao Jiang, Pengfei Liu, Yiwei Wang, Ming-Hsuan Yang
Abstract: Recent coding agents can generate complete codebases from simple prompts, yet existing evaluations focus on issue-level bug fixing and lag behind end-to-end development. We introduce ProjDevBench, an end-to-end benchmark that provides project requirements to coding agents and evaluates the resulting repositories. Combining Online Judge (OJ) testing with LLM-assisted code review, the benchmark evaluates agents on (1) system architecture design, (2) functional correctness, and (3) iterative solution refinement. We curate 20 programming problems across 8 categories, covering both concept-oriented tasks and real-world application scenarios, and evaluate six coding agents built on different LLM backends. Our evaluation reports an overall acceptance rate of 27.38%: agents handle basic functionality and data structures but struggle with complex system design, time complexity optimization, and resource management. Our benchmark is available at https://github.com/zsworld6/projdevbench.
Authors: Mingda Zhang, Haoran Luo, Tiesunlong Shen, Qika Lin, Xiaoying Tang, Rui Mao, Erik Cambria
Abstract: In recent years, a variety of powerful agentic workflows have been applied to solve a wide range of human problems. However, existing workflow orchestration still faces key challenges, including high manual cost, reliance on specific operators/large language models (LLMs), and sparse reward signals. To address these challenges, we propose FlowSteer, an end-to-end reinforcement learning framework that takes a lightweight policy model as the agent and an executable canvas environment, automating workflow orchestration through multi-turn interaction. In this process, the policy model analyzes execution states and selects editing actions, while the canvas executes operators and returns feedback for iterative refinement. Moreover, FlowSteer provides a plug-and-play framework that supports diverse operator libraries and interchangeable LLM backends. To effectively train this interaction paradigm, we propose Canvas Workflow Relative Policy Optimization (CWRPO), which introduces diversity-constrained rewards with conditional release to stabilize learning and suppress shortcut behaviors. Experimental results on twelve datasets show that FlowSteer significantly outperforms baselines across various tasks.
Authors: Qixin Xiao
Abstract: Video language models (Video-LLMs) are prone to hallucinations, often generating plausible but ungrounded content when visual evidence is weak, ambiguous, or biased. Existing decoding methods, such as contrastive decoding (CD), rely on random perturbations to construct contrastive data for mitigating hallucination patterns. However, such a way is hard to control the visual cues that drive hallucination or well align with model weaknesses. We propose Model-aware Counterfactual Data based Contrastive Decoding (MACD), a new inference strategy that combines model-guided counterfactual construction with decoding. Our approach uses the Video-LLM's own feedback to identify object regions most responsible for hallucination, generating targeted counterfactual inputs at the object level rather than arbitrary frame or temporal modifications. These model-aware counterfactual data is then integrated into CD to enforce evidence-grounded token selection during decoding. Experiments on EventHallusion, MVBench, Perception-test and Video-MME show that MACD consistently reduces hallucination while maintaining or improving task accuracy across diverse Video-LLMs, including Qwen and InternVL families. The method is especially effective in challenging scenarios involving small, occluded, or co-occurring objects. Our code and data will be publicly released.
Authors: Guhong Chen, Chenghao Sun, Cheng Fu, Qiyao Wang, Zhihong Huang, Chaopeng Wei, Guangxu Chen, Feiteng Fang, Ahmadreza Argha, Bing Zhao, Xander Xu, Qi Han, Hamid Alinejad-Rokny, Qiang Qu, Binhua Li, Shiwen Ni, Min Yang, Hu Wei, Yongbin Li
Abstract: As code large language models (LLMs) evolve into tool-interactive agents via the Model Context Protocol (MCP), their generalization is increasingly limited by low-quality synthetic data and the diminishing returns of quantity scaling. Moreover, quantity-centric scaling exhibits an early bottleneck that underutilizes trajectory data. We propose TDScaling, a Trajectory Diversity Scaling-based data synthesis framework for code agents that scales performance through diversity rather than raw volume. Under a fixed training budget, increasing trajectory diversity yields larger gains than adding more trajectories, improving the performance-cost trade-off for agent training. TDScaling integrates four innovations: (1) a Business Cluster mechanism that captures real-service logical dependencies; (2) a blueprint-driven multi-agent paradigm that enforces trajectory coherence; (3) an adaptive evolution mechanism that steers synthesis toward long-tail scenarios using Domain Entropy, Reasoning Mode Entropy, and Cumulative Action Complexity to prevent mode collapse; and (4) a sandboxed code tool that mitigates catastrophic forgetting of intrinsic coding capabilities. Experiments on general tool-use benchmarks (BFCL, tau^2-Bench) and code agent tasks (RebenchT, CodeCI, BIRD) demonstrate a win-win outcome: TDScaling improves both tool-use generalization and inherent coding proficiency. We plan to release the full codebase and the synthesized dataset (including 30,000+ tool clusters) upon publication.
Authors: Jianhao Ruan, Zhihao Xu, Yiran Peng, Fashen Ren, Zhaoyang Yu, Xinbing Liang, Jinyu Xiang, Yongru Chen, Bang Liu, Chenglin Wu, Yuyu Luo, Jiayi Zhang
Abstract: Language agents have shown strong promise for task automation. Realizing this promise for increasingly complex, long-horizon tasks has driven the rise of a sub-agent-as-tools paradigm for multi-turn task solving. However, existing designs still lack a dynamic abstraction view of sub-agents, thereby hurting adaptability. We address this challenge with a unified, framework-agnostic agent abstraction that models any agent as a tuple Instruction, Context, Tools, Model. This tuple acts as a compositional recipe for capabilities, enabling the system to spawn specialized executors for each task on demand. Building on this abstraction, we introduce an agentic system AOrchestra, where the central orchestrator concretizes the tuple at each step: it curates task-relevant context, selects tools and models, and delegates execution via on-the-fly automatic agent creation. Such designs enable reducing human engineering efforts, and remain framework-agnostic with plug-and-play support for diverse agents as task executors. It also enables a controllable performance-cost trade-off, allowing the system to approach Pareto-efficient. Across three challenging benchmarks (GAIA, SWE-Bench, Terminal-Bench), AOrchestra achieves 16.28% relative improvement against the strongest baseline when paired with Gemini-3-Flash. The code is available at: https://github.com/FoundationAgents/AOrchestra
Authors: Aditya Basarkar, Benyamin Tabarsi, Tiffany Barnes, Dongkuan Xu
Abstract: Mathematical problem solving is a fundamental benchmark for assessing the reasoning capabilities of artificial intelligence and a gateway to applications in education, science, and engineering where reliable symbolic reasoning is essential. Although recent advances in multi-agent LLM-based systems have enhanced their mathematical reasoning capabilities, they still lack a reliably revisable representation of the reasoning process. Existing agents either operate in rigid sequential pipelines that cannot correct earlier steps or rely on heuristic self-evaluation that can fail to identify and fix errors. In addition, programmatic context can distract language models and degrade accuracy. To address these gaps, we introduce Iteratively Improved Program Construction (IIPC), a reasoning method that iteratively refines programmatic reasoning chains and combines execution feedback with the native Chain-of-thought abilities of the base LLM to maintain high-level contextual focus. IIPC surpasses competing approaches in the majority of reasoning benchmarks on multiple base LLMs. All code and implementations are released as open source.
Authors: Simon Kohaut, Benedict Flade, Julian Eggert, Kristian Kersting, Devendra Singh Dhami
Abstract: Exact inference in complex probabilistic models often incurs prohibitive computational costs. This challenge is particularly acute for autonomous agents in dynamic environments that require frequent, real-time belief updates. Existing methods are often inefficient for ongoing reasoning, as they re-evaluate the entire model upon any change, failing to exploit that real-world information streams have heterogeneous update rates. To address this, we approach the problem from a reactive, asynchronous, probabilistic reasoning perspective. We first introduce Resin (Reactive Signal Inference), a probabilistic programming language that merges probabilistic logic with reactive programming. Furthermore, to provide efficient and exact semantics for Resin, we propose Reactive Circuits (RCs). Formulated as a meta-structure over Algebraic Circuits and asynchronous data streams, RCs are time-dynamic Directed Acyclic Graphs that autonomously adapt themselves based on the volatility of input signals. In high-fidelity drone swarm simulations, our approach achieves several orders of magnitude of speedup over frequency-agnostic inference. We demonstrate that RCs' structural adaptations successfully capture environmental dynamics, significantly reducing latency and facilitating reactive real-time reasoning. By partitioning computations based on the estimated Frequency of Change in the asynchronous inputs, large inference tasks can be decomposed into individually memoized sub-problems. This ensures that only the specific components of a model affected by new information are re-evaluated, drastically reducing redundant computation in streaming contexts.
Authors: Benny Cheung
Abstract: Traditional ontologies describe domain structure but cannot generate novel artifacts. Large language models generate fluently but produce outputs lacking structural validity, hallucinating mechanisms without components, goals without end conditions. We introduce Generative Ontology, a framework synthesizing these complementary strengths: ontology provides the grammar; the LLM provides the creativity. Generative Ontology encodes domain knowledge as executable Pydantic schemas constraining LLM generation via DSPy signatures. A multi-agent pipeline assigns specialized roles: a Mechanics Architect designs game systems, a Theme Weaver integrates narrative, a Balance Critic identifies exploits, each carrying a professional "anxiety" that prevents shallow outputs. Retrieval-augmented generation grounds designs in precedents from existing exemplars. We demonstrate the framework through GameGrammar, generating complete tabletop game designs, and present three empirical studies. An ablation study (120 designs, 4 conditions) shows multi-agent specialization produces the largest quality gains (fun d=1.12, depth d=1.59; p<.001), while schema validation eliminates structural errors (d=4.78). A benchmark against 20 published board games reveals structural parity but a bounded creative gap (fun d=1.86): generated designs score 7-8 while published games score 8-9. A test-retest study (50 evaluations) validates the LLM-based evaluator, with 7/9 metrics achieving Good-to-Excellent reliability (ICC 0.836-0.989). The pattern generalizes beyond games. Any domain with expert vocabulary, validity constraints, and accumulated exemplars is a candidate for Generative Ontology.
Authors: Zihao Jiang, Miao Peng, Zhenyan Shan, Wenjie Xu, Ben Liu, Gong Chen, Ziqi Gao, Min Peng
Abstract: Temporal knowledge graph question answering (TKGQA) aims to answer time-sensitive questions by leveraging temporal knowledge bases. While Large Language Models (LLMs) demonstrate significant potential in TKGQA, current prompting strategies constrain their efficacy in two primary ways. First, they are prone to reasoning hallucinations under complex temporal constraints. Second, static prompting limits model autonomy and generalization, as it lack optimization through dynamic interaction with temporal knowledge graphs (TKGs) environments. To address these limitations, we propose \textbf{TKG-Thinker}, a novel agent equipped with autonomous planning and adaptive retrieval capabilities for reasoning over TKGs. Specifically, TKG-Thinker performs in-depth temporal reasoning through dynamic multi-turn interactions with TKGs via a dual-training strategy. We first apply Supervised Fine-Tuning (SFT) with chain of thought data to instill core planning capabilities, followed by a Reinforcement Learning (RL) stage that leverages multi-dimensional rewards to refine reasoning policies under intricate temporal constraints. Experimental results on benchmark datasets with three open-source LLMs show that TKG-Thinker achieves state-of-the-art performance and exhibits strong generalization across complex TKGQA settings.
Authors: Sindhuja Chaduvula, Jessee Ho, Kina Kim, Aravind Narayanan, Mahshid Alinoori, Muskan Garg, Dhanesh Ramachandram, Shaina Raza
Abstract: Over the last decade, explainable AI has primarily focused on interpreting individual model predictions, producing post-hoc explanations that relate inputs to outputs under a fixed decision structure. Recent advances in large language models (LLMs) have enabled agentic AI systems whose behaviour unfolds over multi-step trajectories. In these settings, success and failure are determined by sequences of decisions rather than a single output. While useful, it remains unclear how explanation approaches designed for static predictions translate to agentic settings where behaviour emerges over time. In this work, we bridge the gap between static and agentic explainability by comparing attribution-based explanations with trace-based diagnostics across both settings. To make this distinction explicit, we empirically compare attribution-based explanations used in static classification tasks with trace-based diagnostics used in agentic benchmarks (TAU-bench Airline and AssistantBench). Our results show that while attribution methods achieve stable feature rankings in static settings (Spearman $\rho = 0.86$), they cannot be applied reliably to diagnose execution-level failures in agentic trajectories. In contrast, trace-grounded rubric evaluation for agentic settings consistently localizes behaviour breakdowns and reveals that state tracking inconsistency is 2.7$\times$ more prevalent in failed runs and reduces success probability by 49\%. These findings motivate a shift towards trajectory-level explainability for agentic systems when evaluating and diagnosing autonomous AI behaviour. Resources: https://github.com/VectorInstitute/unified-xai-evaluation-framework https://vectorinstitute.github.io/unified-xai-evaluation-framework
URLs: https://github.com/VectorInstitute/unified-xai-evaluation-framework, https://vectorinstitute.github.io/unified-xai-evaluation-framework
Authors: Alisia Lupidi, Bhavul Gauri, Thomas Simon Foster, Bassel Al Omari, Despoina Magka, Alberto Pepe, Alexis Audran-Reiss, Muna Aghamelu, Nicolas Baldwin, Lucia Cipolina-Kun, Jean-Christophe Gagnon-Audet, Chee Hau Leow, Sandra Lefdal, Hossam Mossalam, Abhinav Moudgil, Saba Nazir, Emanuel Tewolde, Isabel Urrego, Jordi Armengol Estape, Amar Budhiraja, Gaurav Chaurasia, Abhishek Charnalia, Derek Dunfield, Karen Hambardzumyan, Daniel Izcovich, Martin Josifoski, Ishita Mediratta, Kelvin Niu, Parth Pathak, Michael Shvartsman, Edan Toledo, Anton Protopopov, Roberta Raileanu, Alexander Miller, Tatiana Shavrina, Jakob Foerster, Yoram Bachrach
Abstract: LLM agents hold significant promise for advancing scientific research. To accelerate this progress, we introduce AIRS-Bench (the AI Research Science Benchmark), a suite of 20 tasks sourced from state-of-the-art machine learning papers. These tasks span diverse domains, including language modeling, mathematics, bioinformatics, and time series forecasting. AIRS-Bench tasks assess agentic capabilities over the full research lifecycle -- including idea generation, experiment analysis and iterative refinement -- without providing baseline code. The AIRS-Bench task format is versatile, enabling easy integration of new tasks and rigorous comparison across different agentic frameworks. We establish baselines using frontier models paired with both sequential and parallel scaffolds. Our results show that agents exceed human SOTA in four tasks but fail to match it in sixteen others. Even when agents surpass human benchmarks, they do not reach the theoretical performance ceiling for the underlying tasks. These findings indicate that AIRS-Bench is far from saturated and offers substantial room for improvement. We open-source the AIRS-Bench task definitions and evaluation code to catalyze further development in autonomous scientific research.
Authors: Huang Hu, Xianchao Wu, Bingfeng Luo, Chongyang Tao, Can Xu, Wei Wu, Zhan Chen
Abstract: The 20 Questions (Q20) game is a well known game which encourages deductive reasoning and creativity. In the game, the answerer first thinks of an object such as a famous person or a kind of animal. Then the questioner tries to guess the object by asking 20 questions. In a Q20 game system, the user is considered as the answerer while the system itself acts as the questioner which requires a good strategy of question selection to figure out the correct object and win the game. However, the optimal policy of question selection is hard to be derived due to the complexity and volatility of the game environment. In this paper, we propose a novel policy-based Reinforcement Learning (RL) method, which enables the questioner agent to learn the optimal policy of question selection through continuous interactions with users. To facilitate training, we also propose to use a reward network to estimate the more informative reward. Compared to previous methods, our RL method is robust to noisy answers and does not rely on the Knowledge Base of objects. Experimental results show that our RL method clearly outperforms an entropy-based engineering system and has competitive performance in a noisy-free simulation environment.
Authors: Bowen Peng, Jeffrey Quesnelle, Honglu Fan, Enrico Shippole
Abstract: Rotary Position Embeddings (RoPE) have been shown to effectively encode positional information in transformer-based language models. However, these models fail to generalize past the sequence length they were trained on. We present YaRN (Yet another RoPE extensioN method), a compute-efficient method to extend the context window of such models, requiring 10x less tokens and 2.5x less training steps than previous methods. Using YaRN, we show that LLaMA models can effectively utilize and extrapolate to context lengths much longer than their original pre-training would allow, while also surpassing previous the state-of-the-art at context window extension. In addition, we demonstrate that YaRN exhibits the capability to extrapolate beyond the limited context of a fine-tuning dataset. Code is available at https://github.com/jquesnelle/yarn
Authors: Yueming Lyu, Kang Zhao, Bo Peng, Huafeng Chen, Yue Jiang, Yingya Zhang, Jing Dong, Caifeng Shan
Abstract: Text-guided image editing faces significant challenges when considering training and inference flexibility. Much literature collects large amounts of annotated image-text pairs to train text-conditioned generative models from scratch, which is expensive and not efficient. After that, some approaches that leverage pre-trained vision-language models have been proposed to avoid data collection, but they are limited by either per text-prompt optimization or inference-time hyper-parameters tuning. To address these issues, we investigate and identify a specific space, referred to as CLIP DeltaSpace, where the CLIP visual feature difference of two images is semantically aligned with the CLIP textual feature difference of their corresponding text descriptions. Based on DeltaSpace, we propose a novel framework called DeltaEdit, which maps the CLIP visual feature differences to the latent space directions of a generative model during the training phase, and predicts the latent space directions from the CLIP textual feature differences during the inference phase. And this design endows DeltaEdit with two advantages: (1) text-free training; (2) generalization to various text prompts for zero-shot inference. Extensive experiments validate the effectiveness and versatility of DeltaEdit with different generative models, including both the GAN model and the diffusion model, in achieving flexible text-guided image editing. Code is available at https://github.com/Yueming6568/DeltaEdit.
Authors: Dennis Monari, Farhad Fassihi Tash, Jordan J. Bird, Ahmad Lotfi, Isibor Kennedy Ihianle, Salisu Wada Yahaya, Isibor Kennedy Ihianle, Md Mahmudul Hasan, Pedro Sousa, Pedro Machado
Abstract: Invasive signal crayfish have a detrimental impact on ecosystems. They spread the fungal-type crayfish plague disease (Aphanomyces astaci) that is lethal to the native white clawed crayfish, the only native crayfish species in Britain. Invasive signal crayfish extensively burrow, causing habitat destruction, erosion of river banks and adverse changes in water quality, while also competing with native species for resources leading to declines in native populations. Moreover, pollution exacerbates the vulnerability of White-clawed crayfish, with their populations declining by over 90%. To safeguard aquatic ecosystems, it is imperative to address the challenges posed by invasive species and pollution in aquatic ecosystem's. This article introduces the Cognitive Edge Device (CED) computing platform for the detection of crayfish and plastic. It also presents two publicly available underwater datasets, annotated with sequences of crayfish and aquatic plastic debris. Four You Only Look Once (YOLO) variants were trained and evaluated for crayfish and plastic object detection. YOLOv5s achieved the highest detection accuracy, with an mAP@0.5 of 0.90, and achieved the best precision
Authors: Amin Tabrizian, Zhitong Huang, Arsyi Aziz, Peng Wei
Abstract: Delayed and partially observable state information poses significant challenges for reinforcement learning (RL)-based control in real-world autonomous driving. In highway on-ramp merging, a roadside unit (RSU) can sense nearby traffic, perform edge perception, and transmit state estimates to the ego vehicle over vehicle-to-infrastructure (V2I) links. With recent advancements in intelligent transportation infrastructure and edge computing, such RSU-assisted perception is increasingly realistic and already deployed in modern connected roadway systems. However, edge processing time and wireless transmission can introduce stochastic V2I communication delays, violating the Markov assumption and substantially degrading control performance. In this work, we propose DAROM, a Delay-Aware Reinforcement Learning framework for On-ramp Merging that is robust to stochastic delays. We model the problem as a random delay Markov decision process (RDMDP) and develop a unified RL agent for joint longitudinal and lateral control. To recover a Markovian representation under delayed observations, we introduce a Delay-Aware Encoder that conditions on delayed observations, masked action histories, and observed delay magnitude to infer the current latent state. We further integrate a physics-based safety controller to reduce collision risk during merging. Experiments in the Simulation of Urban MObility (SUMO) simulator using real-world traffic data from the Next Generation Simulation (NGSIM) dataset demonstrate that DAROM consistently outperforms standard RL baselines across traffic densities. In particular, the gated recurrent unit (GRU)-based encoder achieves over 99% success in high-density traffic with random V2I delays of up to 2.0 seconds.
Authors: Davide Frizzo, Francesco Borsatti, Alessio Arcudi, Antonio De Moliner, Roberto Oboe, Gian Antonio Susto
Abstract: Anomaly Detection (AD) is crucial in industrial settings to streamline operations by detecting underlying issues. Conventional methods merely label observations as normal or anomalous, lacking crucial insights. In Industry 5.0, interpretable outcomes become desirable to enable users to understand the rational under model decisions. This paper presents the first industrial application of ExIFFI, a recent approach for fast, efficient explanations for the Extended Isolation Forest (EIF) AD method. ExIFFI is tested on three industrial datasets, demonstrating superior explanation effectiveness, computational efficiency and improved raw anomaly detection performances. ExIFFI reaches over then 90\% of average precision on all the benchmarks considered in the study and overperforms state-of-the-art Explainable Artificial Intelligence (XAI) approaches in terms of the feature selection proxy task metric which was specifically introduced to quantitatively evaluate model explanations.
Authors: Fakrul Islam Tushar, Avivah Wang, Lavsen Dahal, Ehsan Samei, Michael R. Harowicz, Jayashree Kalpathy-Cramer, Kyle J. Lafata, Tina D. Tailor, Cynthia Rudin, Joseph Y. Lo
Abstract: Evaluation of artificial intelligence (AI) models for low-dose CT lung cancer screening is limited by heterogeneous datasets, annotation standards, and evaluation protocols, making performance difficult to compare and translate across clinical settings. We establish a public, reproducible multi-dataset benchmark for lung nodule detection and nodule-level cancer classification and quantify cross-dataset generalizability. Using the Duke Lung Cancer Screening (DLCS) dataset as a clinically curated development set, we evaluate performance across LUNA16/LIDC-IDRI, NLST-3D, and LUNA25. Detection models trained on DLCS and LUNA16 were evaluated externally on NLST-3D using free-response ROC analysis. For malignancy classification, we compared five strategies: randomly initialized ResNet50, Models Genesis, Med3D, a Foundation Model for Cancer Biomarkers, and a Strategic Warm-Start (ResNet50-SWS) approach pretrained using detection-derived candidate patches stratified by confidence. Performance was summarized using AUC with 95% confidence intervals and DeLong tests. Detection performance varied substantially by training dataset, with DLCS-trained models outperforming LUNA16-trained models on external NLST-3D evaluation (sensitivity at 2 false positives per scan: 0.72 vs. 0.64; p < 0.001). For malignancy classification, ResNet50-SWS achieved AUCs of 0.71 (DLCS), 0.90 (LUNA16), 0.81 (NLST-3D), and 0.80 (LUNA25), consistently matching or exceeding alternative pretraining strategies. These results demonstrate that dataset characteristics strongly influence lung cancer AI performance and highlight the need for transparent, multi-dataset benchmarking.
Authors: Julian Sester, C\'ecile Decker
Abstract: We propose a robust Q-learning algorithm for Markov decision processes under model uncertainty when each state-action pair is associated with a finite ambiguity set of candidate transition kernels. This finite-measure framework enables highly flexible, user-designed uncertainty models and goes beyond the common KL and Wasserstein ball formulations. We establish almost sure convergence of the learned Q-function to the robust optimum, and derive non-asymptotic high-probability error bounds that separate stochastic approximation error from transition-kernel estimation error. Finally, we show that Wasserstein ball and parametric ambiguity sets can be approximated by finite ambiguity sets, allowing our algorithm to be used as a generic solver beyond the finite setting.
Authors: Taeyoung Kim, Myungjoo Kang
Abstract: The Rectified Power Unit (RePU) activation function, a differentiable generalization of the Rectified Linear Unit (ReLU), has shown promise in constructing neural networks due to its smoothness properties. However, deep RePU networks often suffer from critical issues such as vanishing or exploding values during training, rendering them unstable regardless of hyperparameter initialization. Leveraging the perspective of effective field theory, we identify the root causes of these failures and propose the Modified Rectified Power Unit (MRePU) activation function. MRePU addresses RePU's limitations while preserving its advantages, such as differentiability and universal approximation properties. Theoretical analysis demonstrates that MRePU satisfies criticality conditions necessary for stable training, placing it in a distinct universality class. Extensive experiments validate the effectiveness of MRePU, showing significant improvements in training stability and performance across various tasks, including polynomial regression, physics-informed neural networks (PINNs) and real-world vision tasks. Our findings highlight the potential of MRePU as a robust alternative for building deep neural networks.
Authors: Keqin Li, Jin Wang, Xubo Wu, Xirui Peng, Runmian Chang, Xiaoyu Deng, Yiwen Kang, Yue Yang, Fanghao Ni, Bo Hong
Abstract: With the rapid growth of global e-commerce, the demand for automation in the logistics industry is increasing. This study focuses on automated picking systems in warehouses, utilizing deep learning and reinforcement learning technologies to enhance picking efficiency and accuracy while reducing system failure rates. Through empirical analysis, we demonstrate the effectiveness of these technologies in improving robot picking performance and adaptability to complex environments. The results show that the integrated machine learning model significantly outperforms traditional methods, effectively addressing the challenges of peak order processing, reducing operational errors, and improving overall logistics efficiency. Additionally, by analyzing environmental factors, this study further optimizes system design to ensure efficient and stable operation under variable conditions. This research not only provides innovative solutions for logistics automation but also offers a theoretical and empirical foundation for future technological development and application.
Authors: Xiangyuan Xue, Zeyu Lu, Di Huang, Zidong Wang, Wanli Ouyang, Lei Bai
Abstract: Much previous AI research has focused on developing monolithic models to maximize their intelligence, with the primary goal of enhancing performance on specific tasks. In contrast, this work attempts to study using LLM-based agents to design collaborative AI systems autonomously. To explore this problem, we first introduce ComfyBench to evaluate agents's ability to design collaborative AI systems in ComfyUI. ComfyBench is a comprehensive benchmark comprising 200 diverse tasks covering various instruction-following generation challenges, along with detailed annotations for 3,205 nodes and 20 workflows. Based on ComfyBench, we further develop ComfyAgent, a novel framework that empowers LLM-based agents to autonomously design collaborative AI systems by generating workflows. ComfyAgent is based on two core concepts. First, it represents workflows with code, which can be reversibly converted into workflows and executed as collaborative systems by the interpreter. Second, it constructs a multi-agent system that cooperates to learn from existing workflows and generate new workflows for a given task. While experimental results demonstrate that ComfyAgent achieves a comparable resolve rate to o1-preview and significantly surpasses other agents on ComfyBench, ComfyAgent has resolved only 15\% of creative tasks. LLM-based agents still have a long way to go in autonomously designing collaborative AI systems. Progress with ComfyBench is paving the way for more intelligent and autonomous collaborative AI systems.
Authors: Atula Tejaswi, Yoonsang Lee, Sujay Sanghavi, Eunsol Choi
Abstract: While in-context learning is well-studied with decoder-only language models (LLMs), its utility for encoder-only models remains underexplored. We study in-context learning for encoder-only models for text retrieval tasks. Can incorporating in-context examples (query-document pairs) to the target query enhance retriever performance? Our approach, RARe, finetunes a pre-trained model with in-context examples whose query is semantically similar to the target query. This approach achieves performance gains of up to +2.72% nDCG across open-domain retrieval datasets (BeIR, RAR-b) compared to using the target query only as an input. In particular, we find RARe exhibits stronger out-of-domain generalization compared to models using queries without in-context examples, similar to what is seen for in-context learning in LLMs. We further provide analysis on the design choices of in-context example augmentation for retrievers and lay the foundation for future work.
Authors: Haoxiang Zhang, Shi Chang, Arthur Leung, Kishanthan Thangarajah, Boyuan Chen, Hanan Lutfiyya, Ahmed E. Hassan
Abstract: The rise of Foundation Models (FMs) like Large Language Models (LLMs) is revolutionizing software development. Despite the impressive prototypes, transforming FMware into production-ready products demands complex engineering across various domains. A critical but overlooked aspect is performance engineering, which aims at ensuring FMware meets performance goals such as throughput and latency to avoid user dissatisfaction and financial loss. Often, performance considerations are an afterthought, leading to costly optimization efforts post-deployment. FMware's high computational resource demands highlight the need for efficient hardware use. Continuous performance engineering is essential to prevent degradation. This paper highlights the significance of Software Performance Engineering (SPE) in FMware, identifying four key challenges: cognitive architecture design (i.e., the structural design that defines how AI components interact, reason, and interface with classical software components), communication protocols, tuning and optimization, and deployment. These challenges are based on literature surveys and experiences from developing an in-house FMware system. We discuss problems, current practices, and innovative paths for the software engineering community.
Authors: Yuang Zhao, Tianyu Li, Jiadong Chen, Shenrong Ye, Fuxin Jiang, Xiaofeng Gao
Abstract: Long-term Time Series Forecasting (LTSF) is crucial across various domains, but complex deep models like Transformers are often prone to overfitting on extended sequences. Linear Fully Connected models have emerged as a powerful alternative, achieving competitive results with fewer parameters. However, their reliance on a single, monolithic weight matrix leads to quadratic parameter redundancy and an entanglement of temporal and frequential properties. To address this, we propose DiPE-Linear, a novel model that disentangles this monolithic mapping into a sequence of specialized, parameter-efficient modules. DiPE-Linear features three core components: Static Frequential Attention to prioritize critical frequencies, Static Time Attention to focus on key time steps, and Independent Frequential Mapping to independently process frequency components. A Low-rank Weight Sharing policy further enhances efficiency for multivariate data. This disentangled architecture collectively reduces parameter complexity from quadratic to linear and computational complexity to log-linear. Experiments on real-world datasets show that DiPE-Linear delivers state-of-the-art performance with significantly fewer parameters, establishing a new and highly efficient baseline for LTSF. Our code is available at https://github.com/wintertee/DiPE-Linear/
Authors: Elizabeth Lui, Rui Sun, Vatsal Shah, Xihan Xiong, Jiahao Sun, Davide Crapis, William Knottenbelt, Zhipeng Wang
Abstract: Centralization enhances the efficiency of Artificial Intelligence (AI) but also introduces critical challenges, including single points of failure, inherent biases, data privacy risks, and scalability limitations. To address these issues, blockchain-based Decentralized Artificial Intelligence (DeAI) has emerged as a promising paradigm that leverages decentralization and transparency to improve the trustworthiness of AI systems. Despite rapid adoption in industry, the academic community lacks a systematic analysis of DeAI's technical foundations, opportunities, and challenges. This work presents the first Systematization of Knowledge (SoK) on DeAI, offering a formal definition, a taxonomy of existing solutions based on the AI lifecycle, and an in-depth investigation of the roles of blockchain in enabling secure and incentive-compatible collaboration. We further review security risks across the DeAI lifecycle and empirically evaluate representative mitigation techniques. Finally, we highlight open research challenges and future directions for advancing blockchain-based DeAI.
Authors: Bowen Peng, Lizhang Chen, Baiyu Su, Jeffrey Quesnelle, Diederik P. Kingma, Qiang Liu
Abstract: Scaling neural network training increasingly depends on synchronous data-parallelism, yet full-precision gradient all-reduce imposes a severe communication bottleneck. We propose Decoupled Momentum Optimization (DeMo), a drop-in replacement for any momentum-based optimizers that significantly reduces the communication bandwidth while maintaining convergence. DeMo (i) decouples local momentum updates, (ii) applies a fast orthonormal transform (e.g., DCT) followed by top-k sparsification, and (iii) reuses the momentum buffer as error feedback via momentum subtraction. This design reduces per-step communication by up to two orders of magnitude with minimal computational overhead. Experiments on 300M and 1B-parameter DeMo language models show DeMo transmits up to 85x less data per GPU than AdamW-DDP while achieving comparable loss and accuracy. DeMo is topology-agnostic and enables training across multi-datacenter or Ethernet-based setups. Code is available at https://github.com/bloc97/DeMo
Authors: Chang-Jin Li, Jiyuan Zhang, Yun Tang, Jian Li
Abstract: Personality assessment through situational judgment tests (SJTs) offers unique advantages over traditional Likert-type self-report scales, yet their development remains labor-intensive, time-consuming, and heavily dependent on subject matter experts. Recent advances in large language models (LLMs) have shown promise for automatic item generation (AIG). Building on these developments, the present study focuses on developing and evaluating a structured and generalizable framework for automatically generating personality SJTs, using GPT-4 and ChatGPT-5 as empirical examples. Three studies were conducted. Study 1 systematically compared the effects of prompt design and temperature settings on the content validity of LLM-generated items to develop an effective and stable LLM-based AIG approach for personality SJT. Results showed that optimized prompts and a temperature of 1.0 achieved the best balance of creativity and accuracy on GPT-4. Study 2 examined the cross-model generalizability and reproducibility of this automated SJT generation approach through multiple rounds. The results showed that the approach consistently produced reproducible and high-quality items on ChatGPT-5. Study 3 evaluated the psychometric properties of LLM-generated SJTs covering five facets of the Big Five personality traits. Results demonstrated satisfactory reliability and validity across most facets, though limitations were observed in the convergent validity of the compliance facet and certain aspects of criterion-related validity. These findings provide robust evidence that the proposed LLM-based AIG approach can produce culturally appropriate and psychometrically sound SJTs with efficiency comparable to or exceeding traditional methods.
Authors: Mehdi Hosseini Chagahi, Niloufar Delfan, Behzad Moshiri, Md. Jalil Piran, Jaber Hatam Parikhan
Abstract: Intracranial hemorrhage (ICH) refers to the leakage or accumulation of blood within the skull, which occurs due to the rupture of blood vessels in or around the brain. If this condition is not diagnosed in a timely manner and appropriately treated, it can lead to serious complications such as decreased consciousness, permanent neurological disabilities, or even death.The primary aim of this study is to detect the occurrence or non-occurrence of ICH, followed by determining the type of subdural hemorrhage (SDH). These tasks are framed as two separate binary classification problems. By adding two layers to the co-scale convolutional attention (CCA) classifier architecture, we introduce a novel approach for ICH detection. In the first layer, after extracting features from different slices of computed tomography (CT) scan images, we combine these features and select the 50 components that capture the highest variance in the data, considering them as informative features. We then assess the discriminative power of these features using the bootstrap forest algorithm, discarding those that lack sufficient discriminative ability between different classes. This algorithm explicitly determines the contribution of each feature to the final prediction, assisting us in developing an explainable AI model. The features feed into a boosting neural network as a latent feature space. In the second layer, we introduce a novel uncertainty-based fuzzy integral operator to fuse information from different CT scan slices. This operator, by accounting for the dependencies between consecutive slices, significantly improves detection accuracy.
Authors: Xiao-Yin Liu, Guotao Li, Xiao-Hu Zhou, Zeng-Guang Hou
Abstract: Offline preference-based reinforcement learning (PbRL) provides an effective way to overcome the challenges of designing reward and the high costs of online interaction. However, since labeling preference needs real-time human feedback, acquiring sufficient preference labels is challenging. To solve this, this paper proposes a offLine prEference-bAsed RL with high Sample Efficiency (LEASE) algorithm, where a learned transition model is leveraged to generate unlabeled preference data. Considering the pretrained reward model may generate incorrect labels for unlabeled data, we design an uncertainty-aware mechanism to ensure the performance of reward model, where only high confidence and low variance data are selected. Moreover, we provide the generalization bound of reward model to analyze the factors influencing reward accuracy, and demonstrate that the policy learned by LEASE has theoretical improvement guarantee. The developed theory is based on state-action pair, which can be easily combined with other offline algorithms. The experimental results show that LEASE can achieve comparable performance to baseline under fewer preference data without online interaction.
Authors: Tzu-Heng Huang, Manjot Bilkhu, John Cooper, Frederic Sala, Javier Movellan
Abstract: Large-scale web-crawled datasets contain noise, bias, and irrelevant information, necessitating data selection techniques. Existing methods depend on hand-crafted heuristics, downstream datasets, or require expensive influence-based computations -- all of which limit scalability and introduce unwanted data dependencies. To address this, we introduce the Mimic Score, a simple and geometry-based data-quality metric that evaluates utility by measuring alignment between a sample's gradients and a target direction induced by a pre-trained reference model. This leverages readily available model weights, avoids needing validation datasets, and incurs minimal computational overheads. Building on this metric, we propose Grad-Mimic, a two-stage framework that re-weights samples online to accelerate training and aggregates sample utilities offline to construct effective data filters. Empirically, we show that using mimic scores to guide training improves data efficiency, accelerates convergence, yields consistent performance gains across six image datasets, and enhances CLIP models with 20.7% fewer training steps. Additionally, mimic score-based filters augment existing filtering techniques, enabling improved CLIP models trained with 4.7 million fewer samples.
Authors: Jialun Cao, Yuk-Kit Chan, Zixuan Ling, Wenxuan Wang, Shuqing Li, Mingwei Liu, Ruixi Qiao, Yuting Han, Chaozheng Wang, Boxi Yu, Pinjia He, Shuai Wang, Zibin Zheng, Michael R. Lyu, Shing-Chi Cheung
Abstract: Code-related benchmarks play a critical role in evaluating large language models (LLMs), yet their quality fundamentally shapes how the community interprets model capabilities. In the past few years, awareness of benchmark quality has grown. Yet, after a decade-scale (2014-2025) survey over 572 code benchmarks, we observed a lag between growing awareness and actual practice. For example, in 2025 alone, the number of benchmarks that ignore code coverage when providing test cases nearly matches the total count accumulated across the previous ten years. In response, we take a clear position: Code benchmarks must prioritize rigor in benchmark construction, reliability in evaluation, and reproducibility in release. To operationalize this position, we introduce a code benchmark guideline HOW2BENCH with 55 checklists. Finally, our further human study also exposed that the current issues not only stem from the significant effort required, but also from a lack of awareness regarding their importance.
Authors: Keqi Han, Yao Su, Lifang He, Liang Zhan, Sergey Plis, Vince Calhoun, Carl Yang
Abstract: Graph deep learning models, a class of AI-driven approaches employing a message aggregation mechanism, have gained popularity for analyzing the functional brain connectome in neuroimaging. However, their actual effectiveness remains unclear. In this study, we re-examine graph deep learning versus classical machine learning models based on four large-scale neuroimaging studies. Surprisingly, we find that the message aggregation mechanism, a hallmark of graph deep learning models, does not help with predictive performance as typically assumed, but rather consistently degrades it. To address this issue, we propose a hybrid model combining a linear model with a graph attention network through dual pathways, achieving robust predictions and enhanced interpretability by revealing both localized and global neural connectivity patterns. Our findings urge caution in adopting complex deep learning models for functional brain connectome analysis, emphasizing the need for rigorous experimental designs to establish tangible performance gains and perhaps more importantly, to pursue improvements in model interpretability.
Authors: Hongye Cao, Sijia Jing, Yanming Wang, Ziyue Peng, Zhixin Bai, Zhe Cao, Meng Fang, Fan Feng, Boyan Wang, Jiaheng Liu, Tianpei Yang, Jing Huo, Yang Gao, Fanyu Meng, Xi Yang, Chao Deng, Junlan Feng
Abstract: With the rapid advancement of Large Language Models (LLMs), the safety of LLMs has been a critical concern requiring precise assessment. Current benchmarks primarily concentrate on single-turn dialogues or a single jailbreak attack method to assess the safety. Additionally, these benchmarks have not taken into account the LLM's capability of identifying and handling unsafe information in detail. To address these issues, we propose a fine-grained benchmark SafeDialBench for evaluating the safety of LLMs across various jailbreak attacks in multi-turn dialogues. Specifically, we design a two-tier hierarchical safety taxonomy that considers 6 safety dimensions and generates more than 4000 multi-turn dialogues in both Chinese and English under 22 dialogue scenarios. We employ 7 jailbreak attack strategies, such as reference attack and purpose reverse, to enhance the dataset quality for dialogue generation. Notably, we construct an innovative assessment framework of LLMs, measuring capabilities in detecting, and handling unsafe information and maintaining consistency when facing jailbreak attacks. Experimental results across 17 LLMs reveal that Yi-34B-Chat and GLM4-9B-Chat demonstrate superior safety performance, while Llama3.1-8B-Instruct and o3-mini exhibit safety vulnerabilities.
Authors: Martina Miliani, Serena Auriemma, Alessandro Bondielli, Emmanuele Chersoni, Lucia Passaro, Irene Sucameli, Alessandro Lenci
Abstract: Large Language Models (LLMs) are increasingly used in tasks requiring interpretive and inferential accuracy. In this paper, we introduce ExpliCa, a new dataset for evaluating LLMs in explicit causal reasoning. ExpliCa uniquely integrates both causal and temporal relations presented in different linguistic orders and explicitly expressed by linguistic connectives. The dataset is enriched with crowdsourced human acceptability ratings. We tested LLMs on ExpliCa through prompting and perplexity-based metrics. We assessed seven commercial and open-source LLMs, revealing that even top models struggle to reach 0.80 accuracy. Interestingly, models tend to confound temporal relations with causal ones, and their performance is also strongly influenced by the linguistic order of the events. Finally, perplexity-based scores and prompting performance are differently affected by model size.
Authors: Tom Wollschl\"ager, Jannes Elstner, Simon Geisler, Vincent Cohen-Addad, Stephan G\"unnemann, Johannes Gasteiger
Abstract: The safety alignment of large language models (LLMs) can be circumvented through adversarially crafted inputs, yet the mechanisms by which these attacks bypass safety barriers remain poorly understood. Prior work suggests that a single refusal direction in the model's activation space determines whether an LLM refuses a request. In this study, we propose a novel gradient-based approach to representation engineering and use it to identify refusal directions. Contrary to prior work, we uncover multiple independent directions and even multi-dimensional concept cones that mediate refusal. Moreover, we show that orthogonality alone does not imply independence under intervention, motivating the notion of representational independence that accounts for both linear and non-linear effects. Using this framework, we identify mechanistically independent refusal directions. We show that refusal mechanisms in LLMs are governed by complex spatial structures and identify functionally independent directions, confirming that multiple distinct mechanisms drive refusal behavior. Our gradient-based approach uncovers these mechanisms and can further serve as a foundation for future work on understanding LLMs.
Authors: Zachary McBride Lazri, Anirudh Nakra, Ivan Brugere, Danial Dervovic, Antigoni Polychroniadou, Furong Huang, Dana Dachman-Soled, Min Wu
Abstract: Algorithmic fairness is often studied in static or single-agent settings, yet many real-world decision-making systems involve multiple interacting entities whose multi-stage actions jointly influence long-term outcomes. Existing fairness methods applied at isolated decision points frequently fail to mitigate disparities that accumulate over time. Although recent work has modeled fairness as a sequential decision-making problem, it typically assumes centralized agents or simplified dynamics, limiting its applicability to complex social systems. We introduce MAFE, a suite of Multi-Agent Fair Environments designed to simulate realistic, modular, and dynamic systems in which fairness emerges from the interplay of multiple agents. We demonstrate MAFEs across three domains -- loan processing, healthcare, and higher education -- that support heterogeneous agents, configurable interventions, and fairness metrics. The environments are open-source and compatible with standard multi-agent reinforcement learning (MARL) libraries, enabling reproducible evaluation of fairness-aware policies. Through extensive experiments on cooperative use cases, we demonstrate how MAFE facilitates the design of equitable multi-agent algorithms and reveals critical trade-offs between fairness, performance, and coordination. MAFE provides a foundation for systematic progress in dynamic, multi-agent fairness research.
Authors: Jungin Kim, Shinwoo Park, Yo-Sub Han
Abstract: Identifying LLM-generated code through watermarking poses a challenge in preserving functional correctness. Previous methods rely on the assumption that watermarking high-entropy tokens effectively maintains output quality. Our analysis reveals a fundamental limitation of this assumption: syntax-critical tokens such as keywords often exhibit the highest entropy, making existing approaches vulnerable to logic corruption. We present STONE, a syntax-aware watermarking method that embeds watermarks only in non-syntactic tokens and preserves code integrity. For rigorous evaluation, we also introduce STEM, a comprehensive metric that balances three critical dimensions: correctness, detectability, and imperceptibility. Across Python, C++, and Java, STONE preserves correctness, sustains strong detectability, and achieves balanced performance with minimal computational overhead. Our implementation is available at https://github.com/inistory/STONE-watermarking.
Authors: Eunkyung Choi, Youngjin Suh, Siun Lee, Hongseok Oh, Juheon Kang, Won Hur, Hun Park, Wonseok Hwang
Abstract: How capable are large language models (LLMs) in the domain of taxation? Although numerous studies have explored the legal domain, research dedicated to taxation remains scarce. Moreover, the datasets used in these studies are either simplified, failing to reflect the real-world complexities, or not released as open-source. To address this gap, we introduce PLAT, a new benchmark designed to assess the ability of LLMs to predict the legitimacy of additional tax penalties. PLAT comprises 300 examples: (1) 100 binary-choice questions, (2) 100 multiple-choice questions, and (3) 100 essay-type questions, all derived from 100 Korean court precedents. PLAT is constructed to evaluate not only LLMs' understanding of tax law but also their performance in legal cases that require complex reasoning beyond straightforward application of statutes. Our systematic experiments with multiple LLMs reveal that (1) their baseline capabilities are limited, especially in cases involving conflicting issues that require a comprehensive understanding (not only of the statutes but also of the taxpayer's circumstances), and (2) LLMs struggle particularly with the "AC" stages of "IRAC" even for advanced reasoning models like o3, which actively employ inference-time scaling. The dataset is publicly available at: https://huggingface.co/collections/sma1-rmarud/plat-predicting-the-legitimacy-of-punitive-additional-tax
Authors: Fenglin Liu, Jinge Wu, Hongjian Zhou, Xiao Gu, Jiayuan Zhu, Jiazhen Pan, Junde Wu, Soheila Molaei, Anshul Thakur, Lei Clifton, Honghan Wu, David A. Clifton
Abstract: Large Language Models (LLMs) achieve competitive results compared to human experts in medical examinations. However, it remains a challenge to apply LLMs to complex clinical decision-making, which requires a deep understanding of medical knowledge and differs from the standardized, exam-style scenarios commonly used in current efforts. A common approach is to fine-tune LLMs for target tasks, which, however, not only requires substantial data and computational resources but also remains prone to generating `hallucinations'. In this work, we present RiskAgent, which synergizes language models with hundreds of validated clinical decision tools supported by evidence-based medicine, to provide generalizable and faithful recommendations. Our experiments show that RiskAgent not only achieves superior performance on a broad range of clinical risk predictions across diverse scenarios and diseases, but also demonstrates robust generalization in tool learning on the external MedCalc-Bench dataset, as well as in medical reasoning and question answering on three representative benchmarks, MedQA, MedMCQA, and MMLU.
Authors: Shwai He, Weilin Cai, Jiayi Huang, Ang Li
Abstract: The Mixture of Experts (MoE) is an effective architecture for scaling large language models by leveraging sparse expert activation to balance performance and efficiency. However, under expert parallelism, MoE suffers from inference inefficiencies due to imbalanced token-to-expert assignment, where underloaded experts complete computations early but must wait for overloaded experts, leading to global delays. We define this phenomenon as the \textbf{\textit{Straggler Effect}}, as the most burdened experts dictate the overall inference latency. To address this, we first propose \textit{\textbf{Capacity-Aware Token Drop}}, which enforces expert capacity limits by discarding excess tokens from overloaded experts, effectively reducing load imbalance with minimal performance impact (e.g., $30\%$ speedup with only $0.9\%$ degradation on OLMoE). Next, given the presence of low-load experts remaining well below the capacity threshold, we introduce \textit{\textbf{Capacity-Aware Expanded Drop}}, which allows tokens to include additional local experts in their candidate set before enforcing strict local capacity constraints, thereby improving load balance and enhancing the utilization of underused experts. Extensive experiments on both language and multimodal MoE models demonstrate the effectiveness of our approach, yielding substantial gains in expert utilization, model performance, and inference efficiency, e.g., applying Expanded Drop to Mixtral-8$\times$7B-Instruct yields a {0.2\%} average performance improvement and a {1.85$\times$} inference speedup. The code is released at: https://github.com/CASE-Lab-UMD/Capacity-Aware-MoE.
Authors: Thanh Linh Nguyen, Dinh Thai Hoang, Diep N. Nguyen, Quoc-Viet Pham
Abstract: Critical learning periods (CLPs) in federated learning (FL) refer to early stages during which low-quality contributions (e.g., sparse training data availability) can permanently impair the performance of the global model owned by the model owner (i.e., a cloud server). However, existing incentive mechanisms exhibit temporal homogeneity, treating all training rounds as equally important, thereby failing to prioritize and attract high-quality contributions during CLPs. This inefficiency is compounded by information asymmetry due to privacy regulations, where the cloud lacks knowledge of client training capabilities, leading to adverse selection. Thus, in this article, we propose a time-aware contract-theoretic incentive framework, named Right Reward Right Time (R3T), to encourage client involvement, especially during CLPs, to maximize the utility of the cloud server. We formulate a cloud utility function that captures the trade-off between the achieved model performance and rewards allocated for clients' contributions, explicitly accounting for client heterogeneity in time and system capabilities, effort, and joining time. Then, we devise a CLP-aware incentive mechanism deriving an optimal contract design that satisfies individual rationality, incentive compatibility, and budget feasibility constraints, motivating rational clients to participate early and contribute efforts. By providing the right reward at the right time, our approach can attract the highest-quality contributions during CLPs. Simulation and proof-of-concept studies show that R3T mitigates information asymmetry, increases cloud utility, and yields superior economic efficiency compared to conventional incentive mechanisms. Our proof-of-concept results demonstrate up to a 47.6% reduction in the total number of clients and up to a 300% improvement in convergence time while achieving competitive test accuracy.
Authors: Mehdi Hosseini Chagahi, Md. Jalil Piran, Niloufar Delfan, Behzad Moshiri, Jaber Hatam Parikhan
Abstract: Intracranial hemorrhage (ICH) is a critical medical emergency caused by the rupture of cerebral blood vessels, leading to internal bleeding within the skull. Accurate and timely classification of hemorrhage subtypes is essential for effective clinical decision-making. To address this challenge, we propose an advanced pyramid vision transformer (PVT)-based model, leveraging its hierarchical attention mechanisms to capture both local and global spatial dependencies in brain CT scans. Instead of processing all extracted features indiscriminately, A SHAP-based feature selection method is employed to identify the most discriminative components, which are then used as a latent feature space to train a boosting neural network, reducing computational complexity. We introduce an entropy-aware aggregation strategy along with a fuzzy integral operator to fuse information across multiple CT slices, ensuring a more comprehensive and reliable scan-level diagnosis by accounting for inter-slice dependencies. Experimental results show that our PVT-based framework significantly outperforms state-of-the-art deep learning architectures in terms of classification accuracy, precision, and robustness. By combining SHAP-driven feature selection, transformer-based modeling, and an entropy-aware fuzzy integral operator for decision fusion, our method offers a scalable and computationally efficient AI-driven solution for automated ICH subtype classification.
Authors: Jinhao Duan, Fei Kong, Hao Cheng, James Diffenderfer, Bhavya Kailkhura, Lichao Sun, Xiaofeng Zhu, Xiaoshuang Shi, Kaidi Xu
Abstract: Object Hallucination (OH) has been acknowledged as one of the major trustworthy challenges in Large Vision-Language Models (LVLMs). Recent advancements in Large Language Models (LLMs) indicate that internal states, such as hidden states, encode the "overall truthfulness" of generated responses. However, it remains under-explored how internal states in LVLMs function and whether they could serve as "per-token" hallucination indicators, which is essential for mitigating OH. In this paper, we first conduct an in-depth exploration of LVLM internal states with OH issues and discover that (1) LVLM internal states are high-specificity per-token indicators of hallucination behaviors. Moreover, (2) different LVLMs encode universal patterns of hallucinations in common latent subspaces, indicating that there exist "generic truthful directions" shared by various LVLMs. Based on these discoveries, we propose Truthful-Guided Pre-Intervention (TruthPrInt) that first learns the truthful direction of LVLM decoding and then applies truthful-guided inference-time intervention during LVLM decoding. We further propose TruthPrInt to enhance both cross-LVLM and cross-data hallucination detection transferability by constructing and aligning hallucination latent subspaces. We evaluate TruthPrInt in extensive experimental settings, including in-domain and out-of-domain scenarios, over popular LVLMs and OH benchmarks. Experimental results indicate that TruthPrInt significantly outperforms state-of-the-art methods. Codes will be available at https://github.com/jinhaoduan/TruthPrInt.
Authors: Samra Irshad, Seungkyu Lee, Nassir Navab, Hong Joo Lee, Seong Tae Kim
Abstract: The presence of adversarial examples in the physical world poses significant challenges to the deployment of Deep Neural Networks in safety-critical applications such as autonomous driving. Most existing methods for crafting physical-world adversarial examples are ad-hoc, relying on temporary modifications like shadows, laser beams, or stickers that are tailored to specific scenarios. In this paper, we introduce a new class of physical-world adversarial examples, AdvWT, which draws inspiration from the naturally occurring phenomenon of `wear and tear', an inherent property of physical objects. Unlike manually crafted perturbations, `wear and tear' emerges organically over time due to environmental degradation, as seen in the gradual deterioration of outdoor signboards. To achieve this, AdvWT follows a two-step approach. First, a GAN-based, unsupervised image-to-image translation network is employed to model these naturally occurring damages, particularly in the context of outdoor signboards. The translation network encodes the characteristics of damaged signs into a latent `damage style code'. In the second step, we introduce adversarial perturbations into the style code, strategically optimizing its transformation process. This manipulation subtly alters the damage style representation, guiding the network to generate adversarial images where the appearance of damages remains perceptually realistic, while simultaneously ensuring their effectiveness in misleading neural networks. Through comprehensive experiments on two traffic sign datasets, we show that AdvWT effectively misleads DNNs in both digital and physical domains. AdvWT achieves an effective attack success rate, greater robustness, and a more natural appearance compared to existing physical-world adversarial examples. Additionally, integrating AdvWT into training enhances a model's generalizability to real-world damaged signs.
Authors: Apurba Pokharel, Ram Dantu, Shakila Zaman, Vinh Quach, Sirisha Talapuru
Abstract: Blockchain consensus mechanisms have relied on algorithms such as Proof-of-Work (PoW) and Proof-of-Stake (PoS) to ensure network functionality and integrity. However, these approaches struggle with adaptability for decision-making where the opinions of each matter rather than reaching an agreement based on honest majority or weighted consensus. This paper introduces a novel deliberation-based consensus mechanism where Large Language Models (LLMs) act as rational agents engaging in structured discussions to reach a unanimous consensus. By leveraging graded consensus and a multi-round deliberation process, our approach ensures unanimous consensus for definitive problems and graded consensus for prioritized decision problems and policies. We provide a formalization of our system and use it to show that the properties of blockchains are maintained, while also addressing the behavior in terms of adversaries, stalled deliberations, and confidence in consensus. Moreover, experimental results demonstrate system feasibility, showcasing convergence, block properties, and accuracy, which enable deliberative decision-making on blockchain networks.
Authors: Yanlin Wang, Kefeng Duan, Dewu Zheng, Ensheng Shi, Fengji Zhang, Yanli Wang, Jiachi Chen, Xilin Liu, Yuchi Ma, Hongyu Zhang, Qianxiang Wang, Zibin Zheng
Abstract: Code intelligence is an emerging domain in software engineering, aiming to improve the effectiveness and efficiency of various code-related tasks. Recent research suggests that incorporating contextual information beyond the basic original task inputs (i.e., source code) can substantially enhance model performance. Such contextual signals may be obtained directly or indirectly from sources such as API documentation or intermediate representations like abstract syntax trees can significantly improve the effectiveness of code intelligence. Despite growing academic interest, there is a lack of systematic analysis of context in code intelligence. To address this gap, we conduct an extensive literature review of 146 relevant studies published between September 2007 and August 2024. Our investigation yields four main contributions. (1) A quantitative analysis of the research landscape, including publication trends, venues, and the explored domains; (2) A novel taxonomy of context types used in code intelligence; (3) A task-oriented analysis investigating context integration strategies across diverse code intelligence tasks; (4) A critical evaluation of evaluation methodologies for context-aware methods. Based on these findings, we identify fundamental challenges in context utilization in current code intelligence systems and propose a research roadmap that outlines key opportunities for future research.
Authors: Azadeh Beiranvand, Seyed Mehdi Vahidipour
Abstract: Text-attributed graphs (TAGs) present unique challenges in representation learning by requiring models to capture both the semantic richness of node-associated texts and the structural dependencies of the graph. While graph neural networks (GNNs) excel at modeling topological information, they lack the capacity to process unstructured text. Conversely, large language models (LLMs) are proficient in text understanding but are typically unaware of graph structure. In this work, we propose BiGTex (Bidirectional Graph Text), a novel architecture that tightly integrates GNNs and LLMs through stacked Graph-Text Fusion Units. Each unit allows for mutual attention between textual and structural representations, enabling information to flow in both directions, text influencing structure and structure guiding textual interpretation. The proposed architecture is trained using parameter-efficient fine-tuning (LoRA), keeping the LLM frozen while adapting to task-specific signals. Extensive experiments on five benchmark datasets demonstrate that BiGTex achieves state-of-the-art performance in node classification and generalizes effectively to link prediction. An ablation study further highlights the importance of soft prompting and bi-directional attention in the model's success.
Authors: Dilip Arumugam, Thomas L. Griffiths
Abstract: A burgeoning area within reinforcement learning (RL) is the design of sequential decision-making agents centered around large language models (LLMs). While autonomous decision-making agents powered by modern LLMs could facilitate numerous real-world applications, such successes demand agents that are capable of data-efficient RL. One key obstacle to achieving data efficiency in RL is exploration, a challenge that we demonstrate many recent proposals for LLM agent designs struggle to contend with. Meanwhile, classic algorithms from the RL literature known to gracefully address exploration require technical machinery that can be challenging to operationalize in purely natural language settings. In this work, rather than relying on finetuning or in-context learning to coax LLMs into implicitly imitating a RL algorithm, we illustrate how LLMs can be used to explicitly implement an existing RL algorithm (Posterior Sampling for Reinforcement Learning) whose capacity for statistically-efficient exploration is already well-studied. We offer empirical results demonstrating how our LLM-based implementation of a known, data-efficient RL algorithm can be considerably more effective in natural language tasks that demand prudent exploration.
Authors: Shubham Vaishnav, Praveen Kumar Donta, Sindri Magn\'usson
Abstract: IoT networks often face conflicting routing goals such as maximizing packet delivery, minimizing delay, and conserving limited battery energy. These priorities can also change dynamically: for example, an emergency alert requires high reliability, while routine monitoring prioritizes energy efficiency to prolong network lifetime. Existing works, including many deep reinforcement learning approaches, are typically centralized and assume static objectives, making them slow to adapt when preferences shift. We propose a dynamic and fully distributed multi-objective Q-learning routing algorithm that learns multiple per-preference Q-tables in parallel and introduces a novel greedy interpolation policy to act near-optimally for unseen preferences without retraining or central coordination. A theoretical analysis further shows that the optimal value function is Lipschitz-continuous in the preference parameter, ensuring that the proposed greedy interpolation policy yields provably near-optimal behavior. Simulations show that our approach adapts in real time to shifting priorities and achieves up to 80-90\% lower energy consumption and more than 2-5x higher cumulative rewards and packet delivery compared to six baseline protocols, under dynamic and distributed settings. Sensitivity analysis across varying preference window lengths confirms that the proposed DPQ framework consistently achieves higher composite reward than all baseline methods, demonstrating robustness to changes in operating conditions.
Authors: Yuxuan Li, Aoi Naito, Hirokazu Shirado
Abstract: Multi-agent systems built on large language models (LLMs) are expected to enhance decision-making by pooling distributed information, yet systematically evaluating this capability has remained challenging. We introduce HiddenBench, a 65-task benchmark grounded in the Hidden Profile paradigm, which isolates collective reasoning under distributed information from individual reasoning ability. Evaluating 15 frontier LLMs, we find that multi-agent LLMs achieve only 30.1% accuracy under distributed information, compared to 80.7% accuracy for single agents given complete information. We trace this gap to a systematic failure mode: agents cannot recognize or act under latent information asymmetry-they fail to reason about what others might know but have not yet expressed, leading to premature convergence on shared evidence while critical distributed facts remain unexplored. These failures persist across prompting strategies, communication depths, and group sizes-and worsen as groups scale. While some models (e.g., Gemini-2.5-Flash/Pro) outperform others, neither model scale nor individual reasoning accuracy reliably predicts collective performance. Our results identify failures in collective information exploration in decision-making as a key limitation of multi-agent LLMs, and provide a theory-grounded, reproducible framework for diagnosing collective reasoning failures.
Authors: Yicong Zhao, King Yeung Tsang, Harshil Vejendla, Haizhou Shi, Zhuohang Li, Zhigang Hua, Qi Xu, Tunyu Zhang, Yi Wang, Ligong Han, Bradley A. Malin, Hao Wang
Abstract: Large Language Models (LLMs) often exhibit misalignment between the quality of their generated responses and the confidence estimates they assign to them. Bayesian treatments, such as marginalizing over a reliable weight posterior or over the space of reasoning traces, provide an effective remedy, but incur substantial computational overhead due to repeated sampling at test time. To enable accurate uncertainty estimation in a single forward pass, we propose a novel distributional distillation framework (Dist2ill) that trains an LLM to produce multiple diverse reasoning paths within one inference pass, while using a lightweight parametric module to approximate empirical confidence scores derived from the sampling distribution. Extensive experiments demonstrate that Dist2ill preserves reasoning diversity and achieves state-of-the-art uncertainty estimation, substantially improving Expected Calibration Error (ECE) and Negative Log-Likelihood (NLL), while remaining computationally efficient.
Authors: Kaustubh Ponkshe, Shaan Shah, Raghav Singhal, Praneeth Vepakomma
Abstract: Large Language Models (LLMs) rely on safety alignment to produce socially acceptable responses. However, this behavior is known to be brittle: further fine-tuning, even on benign or lightly contaminated data, can degrade safety and reintroduce harmful behaviors. A growing body of work suggests that alignment may correspond to identifiable directions in weight space, forming subspaces that could, in principle, be isolated or preserved to defend against misalignment. In this work, we conduct a comprehensive empirical study of this perspective. We examine whether safety-relevant behavior is concentrated in specific linear subspaces, whether it can be separated from general-purpose learning, and whether harmfulness arises from distinguishable patterns in activations. Across both weight and activation spaces, our findings are consistent: subspaces that amplify safe behaviors also amplify useful ones, and prompts with different safety implications activate overlapping representations. Rather than residing in distinct directions, we show that safety is highly entangled with the general learning components of the model. This suggests that subspace-based defenses face fundamental limitations and underscores the need for alternative strategies to preserve safety under continued training. We corroborate these findings with multiple experiments on five open-source LLMs from the Llama and Qwen families. Our code is publicly available at: https://github.com/CERT-Lab/safety-subspaces.
Authors: Raghav Singhal, Kaustubh Ponkshe, Rohit Vartak, Praneeth Vepakomma
Abstract: Large Language Models have demonstrated strong performance across a wide range of tasks, but adapting them efficiently to new domains remains a key challenge. Parameter-Efficient Fine-Tuning (PEFT) methods address this by introducing lightweight, trainable modules while keeping most pre-trained weights fixed. The prevailing approach, LoRA, models updates using a low-rank decomposition, but its expressivity is inherently constrained by the rank. Recent methods like HiRA aim to increase expressivity by incorporating a Hadamard product with the frozen weights, but still rely on the structure of the pre-trained model. We introduce ABBA, a new PEFT architecture that reparameterizes the update as a Hadamard product of two independently learnable low-rank matrices. In contrast to prior work, ABBA fully decouples the update from the pre-trained weights, enabling both components to be optimized freely. This leads to significantly higher expressivity under the same parameter budget, a property we validate through matrix reconstruction experiments. Empirically, ABBA achieves state-of-the-art results on arithmetic and commonsense reasoning benchmarks, consistently outperforming existing PEFT methods by a significant margin across multiple models. Our code is publicly available at: https://github.com/CERT-Lab/abba.
Authors: Yingming Pu, Tao Lin, Hongyu Chen
Abstract: Large Language Model (LLM)-based multi-agent systems (MAS) demonstrate remarkable potential for scientific discovery. Existing approaches, however, often automate scientific discovery using predefined workflows that lack rationality constraints. This often leads to aimless hypothesizing and a failure to consistently link hypotheses with evidence, thereby hindering the systematic reduction of uncertainty. Overcoming these limitations fundamentally requires a principled approach to exploration. We introduce PiFlow, an information-theoretical framework, treating automated scientific discovery as a structured uncertainty reduction problem guided by principles (e.g., scientific laws). Extensive evaluations across three distinct scientific domains demonstrate that PiFlow (I) improves discovery efficiency by 31.18%~41.73% and solution quality by 12.47%~31.72% against state-of-the-art methods, (II) delivers a 5.6x speedup in time-to-solution while reducing token consumption by up to 27% compared to vanilla agents, and (III) serves as a Plug-and-Play module that generalizes on existing agent architecture. Overall, PiFlow establishes a novel paradigm shift in highly efficient agentic scientific discovery, paving the way for more robust and accelerated AI-driven research.
Authors: Sourav Ganguly, Kishan Panaganti, Arnob Ghosh, Adam Wierman
Abstract: Constrained decision-making is essential for designing safe policies in real-world control systems, yet simulated environments often fail to capture real-world adversities. We consider the problem of learning a policy that will maximize the cumulative reward while satisfying a constraint, even when there is a mismatch between the real model and an accessible simulator/nominal model. In particular, we consider the robust constrained Markov decision problem (RCMDP) where an agent needs to maximize the reward and satisfy the constraint against the worst possible stochastic model under the uncertainty set centered around an unknown nominal model. Primal-dual methods, effective for standard constrained MDP (CMDP), are not applicable here because of the lack of the strong duality property. Further, one cannot apply the standard robust value-iteration based approach on the composite value function either as the worst case models may be different for the reward value function and the constraint value function. We propose a novel technique that effectively minimizes the constraint value function--to satisfy the constraints; on the other hand, when all the constraints are satisfied, it can simply maximize the robust reward value function. We prove that such an algorithm finds a policy with at most $\epsilon$ sub-optimality and feasible policy after $O(\epsilon^{-2})$ iterations. In contrast to the state-of-the-art method, we do not need to employ a binary search, thus, we reduce the computation time by at least 4x for smaller value of discount factor ($\gamma$) and by at least 6x for larger value of $\gamma$.
Authors: Chaitanya Amballa, Sattwik Basu, Yu-Lin Wei, Zhijian Yang, Mehmet Ergezer, Romit Roy Choudhury
Abstract: Neural Radiance Fields (NeRFs) have been remarkably successful at synthesizing novel views of 3D scenes by optimizing a volumetric scene function. This scene function models how optical rays bring color information from a 3D object to the camera pixels. Radio frequency (RF) or audio signals can also be viewed as a vehicle for delivering information about the environment to a sensor. However, unlike camera pixels, an RF/audio sensor receives a mixture of signals that contain many environmental reflections (also called "multipath"). Is it still possible to infer the environment using such multipath signals? We show that with redesign, NeRFs can be taught to learn from multipath signals, and thereby "see" the environment. As a grounding application, we aim to infer the indoor floorplan of a home from sparse WiFi measurements made at multiple locations inside the home. Although a difficult inverse problem, our implicitly learnt floorplans look promising, and enables forward applications, such as indoor signal prediction and basic ray tracing.
Authors: Shichang Zhang, Hongzhe Du, Jiaqi W. Ma, Himabindu Lakkaraju
Abstract: Modern AI systems are typically developed through multiple stages-pretraining, fine-tuning rounds, and subsequent adaptation or alignment, where each stage builds on the previous ones and updates the model in distinct ways. This raises a critical question of accountability: when a deployed model succeeds or fails, which stage is responsible, and to what extent? We pose the accountability attribution problem for tracing model behavior back to specific stages of the model development process. To address this challenge, we propose a general framework that answers counterfactual questions about stage effects: how would the model's behavior have changed if the updates from a particular stage had not occurred? Within this framework, we introduce estimators that efficiently quantify stage effects without retraining the model, accounting for both the data and key aspects of model optimization dynamics, including learning rate schedules, momentum, and weight decay. We demonstrate that our approach successfully quantifies the accountability of each stage to the model's behavior. Based on the attribution results, our method can identify and remove spurious correlations learned during image classification and text toxicity detection tasks that were developed across multiple stages. Our approach provides a practical tool for model analysis and represents a significant step toward more accountable AI development.
Authors: Leheng Sheng, Changshuo Shen, Weixiang Zhao, Junfeng Fang, Xiaohao Liu, Zhenkai Liang, Xiang Wang, An Zhang, Tat-Seng Chua
Abstract: As LLMs are increasingly deployed in real-world applications, ensuring their ability to refuse malicious prompts, especially jailbreak attacks, is essential for safe and reliable use. Recently, activation steering has emerged as an effective approach for enhancing LLM safety by adding a refusal direction vector to internal activations of LLMs during inference, which will further induce the refusal behaviors of LLMs. However, indiscriminately applying activation steering fundamentally suffers from the trade-off between safety and utility, since the same steering vector can also lead to over-refusal and degraded performance on benign prompts. Although prior efforts, such as vector calibration and conditional steering, have attempted to mitigate this trade-off, their lack of theoretical grounding limits their robustness and effectiveness. To better address the trade-off between safety and utility, we present a theoretically grounded and empirically effective activation steering method called AlphaSteer. Specifically, it considers activation steering as a learnable process with two principled learning objectives: utility preservation and safety enhancement. For utility preservation, it learns to construct a nearly zero vector for steering benign data, with the null-space constraints. For safety enhancement, it learns to construct a refusal direction vector for steering malicious data, with the help of linear regression. Experiments across multiple jailbreak attacks and utility benchmarks demonstrate the effectiveness of AlphaSteer, which significantly improves the safety of LLMs without compromising general capabilities. Our codes are available at https://github.com/AlphaLab-USTC/AlphaSteer.
Authors: Subba Reddy Oota, Khushbu Pahwa, Prachi Jindal, Satya Sai Srinath Namburi, Maneesh Singh, Tanmoy Chakraborty, Bapi S. Raju, Manish Gupta
Abstract: Recent voxel-wise multimodal brain encoding studies have shown that multimodal large language models (MLLMs) exhibit a higher degree of brain alignment compared to unimodal models. More recently, instruction-tuned multimodal (IT) models have been shown to generate task-specific representations that align strongly with brain activity, yet most prior evaluations focus on unimodal stimuli or non-instruction-tuned models under multimodal stimuli. We still lack a clear understanding of whether instruction-tuning is associated with IT-MLLMs organizing their representations around functional task demands or if they simply reflect surface semantics. To address this, we estimate brain alignment by predicting fMRI responses recorded during naturalistic movie watching (video with audio) from MLLM representations. Using instruction-specific embeddings from six video and two audio IT-MLLMs, across 13 video task instructions, we find that instruction-tuned video MLLMs significantly outperform in-context learning (ICL) multimodal models (~9%), non-instruction-tuned multimodal models (~15%), and unimodal baselines (~20%). Our evaluation of MLLMs across video and audio tasks, and language-guided probing produces distinct task-specific MLLM representations that vary across brain regions. We also find that ICL models show strong semantic organization (r=0.78), while IT models show weak coupling to instruction-text semantics (r=0.14), consistent with task-conditioned subspaces associated with higher brain alignment. These findings are consistent with an association between task-specific instructions and stronger brain-MLLM alignment, and open new avenues for mapping joint information processing in both systems. We make the code publicly available [https://github.com/subbareddy248/mllm_videos].
Authors: Sumanth Manduru, Carlotta Domeniconi
Abstract: The rapid adoption of Small Language Models (SLMs) for resource constrained applications has outpaced our understanding of their ethical and fairness implications. To address this gap, we introduce the Vacuous Neutrality Framework (VaNeu), a multi-dimensional evaluation paradigm designed to assess SLM fairness prior to deployment. The framework examines model robustness across four stages - biases, utility, ambiguity handling, and positional bias over diverse social bias categories. To the best of our knowledge, this work presents the first large-scale audit of SLMs in the 0.5-5B parameter range, an overlooked "middle tier" between BERT-class encoders and flagship LLMs. We evaluate nine widely used SLMs spanning four model families under both ambiguous and disambiguated contexts. Our findings show that models demonstrating low bias in early stages often fail subsequent evaluations, revealing hidden vulnerabilities and unreliable reasoning. These results underscore the need for a more comprehensive understanding of fairness and reliability in SLMs, and position the proposed framework as a principled tool for responsible deployment in socially sensitive settings.
Authors: Antonin Berthon, Mihaela van der Schaar
Abstract: Accurately assessing student knowledge is central to education. Cognitive Diagnosis (CD) models estimate student proficiency at a fixed point in time, while Knowledge Tracing (KT) methods model evolving knowledge states to predict future performance. However, existing approaches either provide quantitative concept mastery estimates with limited expressivity (CD, probabilistic KT) or prioritize predictive accuracy at the cost of interpretability (deep learning KT). We propose Language Bottleneck Models (LBMs), where an encoder LLM produces textual knowledge state summaries, which a decoder LLM uses to predict future performance. This produces interpretable summaries that can express nuanced insights--such as misconceptions--that CD and KT models cannot capture. Extensive validation across synthetic and real-world datasets shows LBMs reveal qualitative insights beyond what CD and KT models can capture, while achieving competitive accuracy with improved sample efficiency. We demonstrate that the encoder and decoder can be fine-tuned with reinforcement learning and supervised fine-tuning respectively to improve both summary quality and predictive performance.
Authors: Xingyu Alice Yang, Jianyu Zhang, L\'eon Bottou
Abstract: Transfer learning is widely used to adapt large pretrained models to new tasks with only a small amount of new data. However, a challenge persists -- the features from the original task often do not fully cover what is needed for unseen data, especially when the relatedness of tasks is not clear. Since deep learning models tend to learn very sparse representations, they retain only the minimal features required for the initial training while discarding potentially ones for downstream transfer. A theoretical framework developed in this work demonstrates that such pretraining captures inconsistent aspects of the data distribution, therefore, inducing transfer bias. To address this limitation, we propose an inexpensive ensembling strategy that aggregates multiple models to generate richer feature representations. On ResNet, this approach yields a $9\%$ improvement in transfer accuracy without incurring extra pretraining cost. We also present empirical evidence from a range of deep learning studies, confirming that the phenomenon is pervasive across modern deep learning architectures. These results suggests that relying solely on large pretrained networks is not always the most effective way to improve model generalization. Instead, fostering richer, more diverse representations -- e.g. - through model ensembles -- can substantially enhance transfer learning performance.
Authors: Xiaona Zhou, Constantin Brif, Ismini Lourentzou
Abstract: Anomaly detection in multivariate time series is essential across domains such as healthcare, cybersecurity, and industrial monitoring, yet remains fundamentally challenging due to high-dimensional dependencies, the presence of cross-correlations between time-dependent variables, and the scarcity of labeled anomalies. We introduce mTSBench, the largest benchmark to date for multivariate time series anomaly detection and model selection, consisting of 344 labeled time series across 19 datasets from a wide range of application domains. We comprehensively evaluate 24 anomaly detectors, including the only two publicly available large language model-based methods for multivariate time series. Consistent with prior findings, we observe that no single detector dominates across datasets, motivating the need for effective model selection. We benchmark three recent model selection methods and find that even the strongest of them remain far from optimal. Our results highlight the outstanding need for robust, generalizable selection strategies. We open-source the benchmark at https://plan-lab.github.io/mtsbench to encourage future research.
Authors: Yifan Sun, Yushan Liang, Zhen Zhang, Xin Liu, Jiaye Teng
Abstract: Self-improvement is a significant techniques within the realm of large language model (LLM), aiming to enhance the LLM performance without relying on external data. Despite its significance, generally how LLM performances evolve during the self-improvement process remains underexplored. In this paper, we theoretically model the training dynamics of self-improvement via the concept of solver-verifier gap. This is inspired by the conjecture that the performance enhancement of self-improvement stems from the gap between LLM's solver capability and verifier capability. Based on the theoretical framework, we further show how to model the entire training trajectory. This framework allows quantifying the capability limit of self-improvement by fitting the theoretical model to the experiment results. We validate the effectiveness of the theoretical framework on various LLMs and datasets. Beyond self-improvement, we extend our analysis to investigate how external data influences these dynamics within the framework. Notably, we find that under limited external data regimes, such external data can be utilized at any stage without significantly affecting final performances, which accords with the empirical observations.
Authors: Shirley Wu, Parth Sarthi, Shiyu Zhao, Aaron Lee, Herumb Shandilya, Adrian Mladenic Grobelnik, Nurendra Choudhary, Eddie Huang, Karthik Subbian, Linjun Zhang, Diyi Yang, James Zou, Jure Leskovec
Abstract: Compound AI systems integrating multiple components, such as Large Language Models, specialized tools, and traditional machine learning models, are increasingly deployed to solve complex real-world tasks. However, optimizing compound systems remains challenging due to their non-differentiable structures and diverse configuration types across components, including prompts, hyperparameters, and model parameters. To address this challenge, we propose Optimas, a unified framework for effective optimization of compound systems. The core idea of Optimas is to maintain one Local Reward Function (LRF) per component, each satisfying a local-global alignment property, i.e., each component's local reward correlates with the global system performance. In each iteration, Optimas efficiently adapts the LRFs to maintain this property while simultaneously maximizing each component's local reward. This approach enables independent updates of heterogeneous configurations using the designated optimization method, while ensuring that local improvements consistently lead to performance gains. We present extensive evaluations across five real-world compound systems to demonstrate that Optimas outperforms strong baselines by an average improvement of 11.92%, offering a general and effective approach for improving compound systems. Our website is at https://optimas.stanford.edu.
Authors: Asad Aali, Vasiliki Bikia, Maya Varma, Nicole Chiou, Sophie Ostmeier, Arnav Singhvi, Magdalini Paschali, Ashwin Kumar, Andrew Johnston, Karimar Amador-Martinez, Eduardo Juan Perez Guerrero, Paola Naovi Cruz Rivera, Sergios Gatidis, Christian Bluethgen, Eduardo Pontes Reis, Eddy D. Zandee van Rilland, Poonam Laxmappa Hosamani, Kevin R Keet, Minjoung Go, Evelyn Ling, David B. Larson, Curtis Langlotz, Roxana Daneshjou, Jason Hom, Sanmi Koyejo, Emily Alsentzer, Akshay S. Chaudhari
Abstract: With the growing use of language models (LMs) in clinical environments, there is an immediate need to evaluate the accuracy and safety of LM-generated medical text. Currently, such evaluation relies solely on manual physician review. However, detecting errors in LM-generated text is challenging because 1) manual review is costly and 2) expert-composed reference outputs are often unavailable in real-world settings. While the "LLM-as-a-judge" paradigm offers scalable evaluation, even frontier LMs can miss subtle but clinically significant errors. We propose MedVAL, a novel, self-supervised, data-efficient distillation method that leverages synthetic data to train evaluator LMs to assess whether LM-generated medical outputs are factually consistent with inputs, without requiring physician labels or reference outputs. To evaluate LM performance, we introduce MedVAL-Bench, a dataset of 840 physician-annotated outputs across 6 diverse medical tasks capturing real-world challenges. Across 10 state-of-the-art LMs spanning open-source and proprietary models, MedVAL distillation significantly improves (p < 0.001) alignment with physicians across seen and unseen tasks, increasing average F1 scores from 66% to 83%. Despite strong baseline performance, MedVAL improves the best-performing proprietary LM (GPT-4o) by 8% without training on physician-labeled data, demonstrating a performance statistically non-inferior to a single human expert on a subset annotated by multiple physicians (p < 0.001). To support a scalable, risk-aware pathway towards clinical integration, we open-source: 1) Codebase (https://github.com/StanfordMIMI/MedVAL), 2) MedVAL-Bench (https://huggingface.co/datasets/stanfordmimi/MedVAL-Bench), 3) MedVAL-4B (https://huggingface.co/stanfordmimi/MedVAL-4B). Our benchmark provides evidence of LMs approaching expert-level ability in validating AI-generated medical text.
URLs: https://github.com/StanfordMIMI/MedVAL),, https://huggingface.co/datasets/stanfordmimi/MedVAL-Bench),, https://huggingface.co/stanfordmimi/MedVAL-4B).
Authors: Tarun G, Naman Malpani, Gugan Thoppe, Sridharan Devarajan
Abstract: Deep neural networks are increasingly applied in automated histopathology. Yet, whole-slide images (WSIs) are often acquired at gigapixel sizes, rendering them computationally infeasible to analyze entirely at high resolution. Diagnostic labels are largely available only at the slide-level, because expert annotation of images at a finer (patch) level is both laborious and expensive. Moreover, regions with diagnostic information typically occupy only a small fraction of the WSI, making it inefficient to examine the entire slide at full resolution. Here, we propose SASHA -- Sequential Attention-based Sampling for Histopathological Analysis -- a deep reinforcement learning approach for efficient analysis of histopathological images. First, SASHA learns informative features with a lightweight hierarchical, attention-based multiple instance learning (MIL) model. Second, SASHA samples intelligently and zooms selectively into a small fraction (10-20\%) of high-resolution patches to achieve reliable diagnoses. We show that SASHA matches state-of-the-art methods that analyze the WSI fully at high resolution, albeit at a fraction of their computational and memory costs. In addition, it significantly outperforms competing, sparse sampling methods. We propose SASHA as an intelligent sampling model for medical imaging challenges that involve automated diagnosis with exceptionally large images containing sparsely informative features. Model implementation is available at: https://github.com/coglabiisc/SASHA.
Authors: Fedor Chernogorskii, Sergei Averkiev, Liliya Kudraleeva, Zaven Martirosian, Maria Tikhonova, Valentin Malykh, Alena Fenogenova
Abstract: This paper introduces DRAGOn, method to design a RAG benchmark on a regularly updated corpus. It features recent reference datasets, a question generation framework, an automatic evaluation pipeline, and a public leaderboard. Specified reference datasets allow for uniform comparison of RAG systems, while newly generated dataset versions mitigate data leakage and ensure that all models are evaluated on unseen, comparable data. The pipeline for automatic question generation extracts the Knowledge Graph from the text corpus and produces multiple question-answer pairs utilizing modern LLM capabilities. A set of diverse LLM-as-Judge metrics is provided for a comprehensive model evaluation. We used Russian news outlets to form the datasets and demonstrate our methodology. We launch a public leaderboard to track the development of RAG systems and encourage community participation.
Authors: Wuxin Wang, Weicheng Ni, Lilan Huang, Tao Hao, Ben Fei, Shuo Ma, Taikang Yuan, Yanlai Zhao, Kefeng Deng, Xiaoyong Li, Hongze Leng, Boheng Duan, Lei Bai, Weimin Zhang, Junqiang Song, Kaijun Ren
Abstract: Machine Learning (ML) has shown great promise in revolutionizing weather forecasting, yet most ML systems still rely on initial conditions generated by Numerical Weather Prediction (NWP) systems. End-to-end ML models aim to eliminate this dependency, but they often rely on observation-specific encoders and require redesign or retraining when observation sources change, thereby limiting their operational robustness. Here, we introduce XiChen, a global weather observation-to-forecast ML system via four-dimensional variational (4DVar) gradient-guided flexible assimilation. We demonstrate that the gradient of the 4DVar cost function serves as a physically grounded interface that maps heterogeneous observations into a common state space. This novel formulation enables XiChen to flexibly assimilate diverse conventional and raw satellite observations while preserving physical consistency. Experiments show that the system achieves forecasting metrics competitive with operational NWP systems. This work provides a practical and physically consistent route toward operational ML-based global weather forecasting systems with heterogeneous and evolving observations.
Authors: Yuji Kanagawa, Kenji Doya
Abstract: Fear is a critical brain function that enables us to learn to avoid danger via reinforcement learning (RL). While many researchers have argued that fear has evolved to escape predators, how varying predatory pressures have shaped fear and other rewards, including positive social rewards for collective grouping, remains an open question. In this study, we investigate the relationship between predatory pressure and fear using an evolutionary simulation of RL agents with evolving rewards. In our simulation, prey and predator RL agents co-evolve their reward functions, including visual rewards for observing prey and predators. While fear-like negative visual rewards for predators often evolved in prey, we also observed cases in which positive rewards for both predators and prey evolved, the latter serving as a social reward for collective grouping. A comparison between different environmental conditions revealed that stronger predator hunting capability promoted stronger fear reward, while less food supply promoted more negative social reward. Moreover, fear did not evolve in response to static pitfalls with non-lethal damage, suggesting that actively hunting predators played an important role in its evolution. These results highlight the special role of predators in the diverse evolution of fear and social rewards.
Authors: Dzung Dinh, Boqi Chen, Yunni Qu, Marc Niethammer, Junier Oliva
Abstract: In many critical domains, features are not freely available at inference time: each measurement may come with a cost of time, money, and risk. Longitudinal prediction further complicates this setting because both features and labels evolve over time, and missing measurements at earlier timepoints may become permanently unavailable. We propose NOCTA, a Non-Greedy Objective Cost-Tradeoff Acquisition framework that sequentially acquires the most informative features at inference time while accounting for both temporal dynamics and acquisition cost. NOCTA is driven by a novel objective, NOCT, which evaluates a candidate set of future feature-time acquisitions by its expected predictive loss together with its acquisition cost. Since NOCT depends on unobserved future trajectories at inference time, we develop two complementary estimators: (i) NOCT-Contrastive, which learns an embedding of partial observations utilizing the induced distribution over future acquisitions, and (ii) NOCT-Amortized, which directly predicts NOCT for candidate plans with a neural network. Experiments on synthetic and real-world medical datasets demonstrate that both NOCTA estimators outperform existing baselines, achieving higher accuracy at lower acquisition costs.
Authors: Jing Gu, Xian Liu, Yu Zeng, Ashwin Nagarajan, Fangrui Zhu, Daniel Hong, Yue Fan, Qianqi Yan, Kaiwen Zhou, Ming-Yu Liu, Xin Eric Wang
Abstract: Video generation models have achieved remarkable progress in creating high-quality, photorealistic content. However, their ability to accurately simulate physical phenomena remains a critical and unresolved challenge. This paper presents PhyWorldBench, a comprehensive benchmark designed to evaluate video generation models based on their adherence to the laws of physics. The benchmark covers multiple levels of physical phenomena, ranging from fundamental principles such as object motion and energy conservation to more complex scenarios involving rigid body interactions and human or animal motion. Additionally, we introduce a novel Anti-Physics category, where prompts intentionally violate real-world physics, enabling the assessment of whether models can follow such instructions while maintaining logical consistency. Besides large-scale human evaluation, we also design a simple yet effective method that utilizes current multimodal large language models to evaluate physics realism in a zero-shot fashion. We evaluate 12 state-of-the-art text-to-video generation models, including five open-source and five proprietary models, with detailed comparison and analysis. Through systematic testing across 1050 curated prompts spanning fundamental, composite, and anti-physics scenarios, we identify pivotal challenges these models face in adhering to real-world physics. We further examine their performance under diverse physical phenomena and prompt types, and derive targeted recommendations for crafting prompts that enhance fidelity to physical principles.
Authors: Jiaji Zhang, Ruichao Sun, Hailiang Zhao, Jiaju Wu, Peng Chen, Hao Li, Yuying Liu, Kingsum Chow, Gang Xiong, Shuiguang Deng
Abstract: Diffusion models have demonstrated exceptional generative capabilities but are computationally intensive, posing significant challenges for deployment in resource-constrained or latency-sensitive environments. Quantization offers an effective means to reduce model size and computational cost, with post-training quantization (PTQ) being particularly appealing due to its compatibility with pre-trained models without requiring retraining or training data. However, existing PTQ methods for diffusion models often rely on architecture-specific heuristics that limit their generalizability and hinder integration with industrial deployment pipelines. To address these limitations, we propose SegQuant, a unified quantization framework that adaptively combines complementary techniques to enhance cross-model versatility. SegQuant consists of a segment-aware, graph-based quantization strategy (SegLinear) that captures structural semantics and spatial heterogeneity, along with a dual-scale quantization scheme (DualScale) that preserves polarity-asymmetric activations, which is crucial for maintaining visual fidelity in generated outputs. SegQuant is broadly applicable beyond Transformer-based diffusion models, achieving strong performance while ensuring seamless compatibility with mainstream deployment tools.
Authors: Yutao Sun, Zhenyu Li, Yike Zhang, Tengyu Pan, Bowen Dong, Yuyi Guo, Jianyong Wang
Abstract: Transformer-based architectures have become the prevailing backbone of large language models. However, the quadratic time and memory complexity of self-attention remains a fundamental obstacle to efficient long-context modeling. To address this limitation, recent research has introduced two principal categories of efficient attention mechanisms. Linear attention methods achieve linear complexity through kernel approximations, recurrent formulations, or fastweight dynamics, thereby enabling scalable inference with reduced computational overhead. Sparse attention techniques, in contrast, limit attention computation to selected subsets of tokens based on fixed patterns, block-wise routing, or clustering strategies, enhancing efficiency while preserving contextual coverage. This survey provides a systematic and comprehensive overview of these developments, integrating both algorithmic innovations and hardware-level considerations. In addition, we analyze the incorporation of efficient attention into largescale pre-trained language models, including both architectures built entirely on efficient attention and hybrid designs that combine local and global components. By aligning theoretical foundations with practical deployment strategies, this work aims to serve as a foundational reference for advancing the design of scalable and efficient language models.
Authors: Sudip K. Seal, Maksudul Alam, Jorge Ramirez, Sajal Dash, Hao Lu
Abstract: Energy efficiency of training and inferencing with large neural network models is a critical challenge facing the future of sustainable large-scale machine learning workloads. This paper introduces an alternative strategy, called phantom parallelism, to minimize the net energy consumption of traditional tensor (model) parallelism, the most energy-inefficient component of large neural network training. The approach is presented in the context of feed-forward network architectures as a preliminary, but comprehensive, proof-of-principle study of the proposed methodology. We derive new forward and backward propagation operators for phantom parallelism, implement them as custom autograd operations within an end-to-end phantom parallel training pipeline and compare its parallel performance and energy-efficiency against those of conventional tensor parallel training pipelines. Formal analyses that predict lower bandwidth and FLOP counts are presented with supporting empirical results on up to 256 GPUs that corroborate these gains. Experiments are shown to deliver approximately 50% reduction in the energy consumed to train FFNs using the proposed phantom parallel approach when compared with conventional tensor parallel methods. Additionally, the proposed approach is shown to train smaller phantom models to the same model loss on smaller GPU counts as larger tensor parallel models on larger GPU counts offering the possibility for even greater energy savings.
Authors: Jun Qi, Chao-Han Huck Yang, Pin-Yu Chen, Min-Hsiu Hsieh
Abstract: Variational Quantum Computing (VQC) faces fundamental scalability barriers, primarily due to barren plateaus and sensitivity to quantum noise. To address these challenges, we introduce TensorHyper-VQC, a novel tensor-train (TT)-guided hypernetwork framework that significantly improves the robustness and scalability of VQC. Our framework fully delegates the generation of quantum-circuit parameters to a classical TT network, thereby decoupling optimization from quantum hardware. This innovative parameterization mitigates gradient vanishing, enhances noise resilience through structured low-rank representations, and facilitates efficient gradient propagation. Grounded in Neural Tangent Kernel and statistical learning theory, our rigorous theoretical analyses establish strong guarantees on approximation capability, optimization stability, and generalization performance. Extensive empirical results across quantum dot classification, Max-Cut optimization, and molecular quantum simulation tasks demonstrate that TensorHyper-VQC consistently achieves superior performance and robust noise tolerance, including hardware-level validation on a 156-qubit IBM Heron processor. These results position TensorHyper-VQC as a scalable and noise-resilient framework for advancing practical quantum machine learning on near-term devices.
Authors: Shunlei Li, Longsen Gao, Jin Wang, Chang Che, Xi Xiao, Jiuwen Cao, Yingbai Hu, Hamid Reza Karimi
Abstract: Teaching robots dexterous skills from human videos remains challenging due to the reliance on low-level trajectory imitation, which fails to generalize across object types, spatial layouts, and manipulator configurations. We propose Graph-Fused Vision-Language-Action (GF-VLA), a framework that enables dual-arm robotic systems to perform task-level reasoning and execution directly from RGB and Depth human demonstrations. GF-VLA first extracts Shannon-information-based cues to identify hands and objects with the highest task relevance, then encodes these cues into temporally ordered scene graphs that capture both hand-object and object-object interactions. These graphs are fused with a language-conditioned transformer that generates hierarchical behavior trees and interpretable Cartesian motion commands. To improve execution efficiency in bimanual settings, we further introduce a cross-hand selection policy that infers optimal gripper assignment without explicit geometric reasoning. We evaluate GF-VLA on four structured dual-arm block assembly tasks involving symbolic shape construction and spatial generalization. Experimental results show that the information-theoretic scene representation achieves over 95 percent graph accuracy and 93 percent subtask segmentation, supporting the LLM planner in generating reliable and human-readable task policies. When executed by the dual-arm robot, these policies yield 94 percent grasp success, 89 percent placement accuracy, and 90 percent overall task success across stacking, letter-building, and geometric reconfiguration scenarios, demonstrating strong generalization and robustness across diverse spatial and semantic variations.
Authors: Danfeng Hong, Chenyu Li, Naoto Yokoya, Bing Zhang, Xiuping Jia, Antonio Plaza, Paolo Gamba, Jon Atli Benediktsson, Jocelyn Chanussot
Abstract: Hyperspectral imaging (HSI) is an advanced sensing modality that simultaneously captures spatial and spectral information, enabling non-invasive, label-free analysis of material, chemical, and biological properties. This Primer presents a comprehensive overview of HSI, from the underlying physical principles and sensor architectures to key steps in data acquisition, calibration, and correction. We summarize common data structures and highlight classical and modern analysis methods, including dimensionality reduction, classification, spectral unmixing, and AI-driven techniques such as deep learning. Representative applications across Earth observation, precision agriculture, biomedicine, industrial inspection, cultural heritage, and security are also discussed, emphasizing HSI's ability to uncover sub-visual features for advanced monitoring, diagnostics, and decision-making. Persistent challenges, such as hardware trade-offs, acquisition variability, and the complexity of high-dimensional data, are examined alongside emerging solutions, including computational imaging, physics-informed modeling, cross-modal fusion, and self-supervised learning. Best practices for dataset sharing, reproducibility, and metadata documentation are further highlighted to support transparency and reuse. Looking ahead, we explore future directions toward scalable, real-time, and embedded HSI systems, driven by sensor miniaturization, self-supervised learning, and foundation models. As HSI evolves into a general-purpose, cross-disciplinary platform, it holds promise for transformative applications in science, technology, and society.
Authors: Ryan Mioduski
Abstract: Virginia's seventeenth- and eighteenth-century land patents survive primarily as narrative metes-and-bounds descriptions, limiting spatial analysis. This study systematically evaluates current-generation large language models (LLMs) in converting these prose abstracts into geographically accurate latitude/longitude coordinates within a focused evaluation context. A digitized corpus of 5,471 Virginia patent abstracts (1695-1732) is released, with 43 rigorously verified test cases serving as an initial, geographically focused benchmark. Six OpenAI models across three architectures-o-series, GPT-4-class, and GPT-3.5-were tested under two paradigms: direct-to-coordinate and tool-augmented chain-of-thought invoking external geocoding APIs. Results were compared against a GIS analyst baseline, Stanford NER geoparser, Mordecai-3 neural geoparser, and a county-centroid heuristic. The top single-call model, o3-2025-04-16, achieved a mean error of 23 km (median 14 km), outperforming the median LLM (37.4 km) by 37.5%, the weakest LLM (50.3 km) by 53.5%, and external baselines by 67% (GIS analyst) and 70% (Stanford NER). A five-call ensemble further reduced errors to 19.2 km (median 12.2 km) at minimal additional cost (~USD 0.20 per grant), outperforming the median LLM by 48.7%. A patentee-name redaction ablation slightly increased error (~7%), showing reliance on textual landmark and adjacency descriptions rather than memorization. The cost-effective gpt-4o-2024-08-06 model maintained a 28 km mean error at USD 1.09 per 1,000 grants, establishing a strong cost-accuracy benchmark. External geocoding tools offer no measurable benefit in this evaluation. These findings demonstrate LLMs' potential for scalable, accurate, cost-effective historical georeferencing.
Authors: Yisu Liu, Chenxing Li, Wanqian Zhang, Wenfu Wang, Meng Yu, Ruibo Fu, Zheng Lin, Weiping Wang, Dong Yu
Abstract: Controllable text-to-audio generation aims to synthesize audio from textual descriptions while satisfying user-specified constraints, including event types, temporal sequences, and onset and offset timestamps. This enables precise control over both the content and temporal structure of the generated audio. Despite recent progress, existing methods still face inherent trade-offs among accurate temporal localization, open-vocabulary scalability, and practical efficiency. To address these challenges, we propose DegDiT, a novel dynamic event graph-guided diffusion transformer framework for open-vocabulary controllable audio generation. DegDiT encodes the events in the description as structured dynamic graphs. The nodes in each graph are designed to represent three aspects: semantic features, temporal attributes, and inter-event connections. A graph transformer is employed to integrate these nodes and produce contextualized event embeddings that serve as guidance for the diffusion model. To ensure high-quality and diverse training data, we introduce a quality-balanced data selection pipeline that combines hierarchical event annotation with multi-criteria quality scoring, resulting in a curated dataset with semantic diversity. Furthermore, we present consensus preference optimization, facilitating audio generation through consensus among multiple reward signals. Extensive experiments on AudioCondition, DESED, and AudioTime datasets demonstrate that DegDiT achieves state-of-the-art performances across a variety of objective and subjective evaluation metrics.
Authors: Yishun Lu, Wesley Armour
Abstract: Modern GPUs are equipped with large amounts of high-bandwidth memory, enabling them to support mini-batch sizes of up to tens of thousands of training samples. However, most existing optimizers struggle to perform effectively at such a large batch size. As batch size increases, gradient noise decreases due to averaging over many samples, limiting the ability of first-order methods to escape sharp or suboptimal minima and reach the global minimum. Meanwhile, second-order methods like the natural gradient with Kronecker-Factored Approximate Curvature (KFAC) often require excessively high damping to remain stable at large batch sizes. This high damping effectively washes out the curvature information that gives these methods their advantage, reducing their performance to that of simple gradient descent. In this paper, we introduce Fisher-Orthogonal Projection (FOP), a novel technique that restores the effectiveness of the second-order method at very large batch sizes, enabling scalable training with improved generalization and faster convergence. FOP constructs a variance-aware update direction by leveraging gradients from two sub-batches, enhancing the average gradient with a component of the gradient difference that is orthogonal to the average under the Fisher-metric.
Authors: Can Jin, Yang Zhou, Qixin Zhang, Hongwu Peng, Di Zhang, Zihan Dong, Marco Pavone, Ligong Han, Zhang-Wei Hong, Tong Che, Dimitris N. Metaxas
Abstract: Test-time scaling (TTS) for large language models (LLMs) has thus far fallen into two largely separate paradigms: (1) reinforcement learning (RL) methods that optimize sparse outcome-based rewards, yet suffer from instability and low sample efficiency; and (2) search-based techniques guided by independently trained, static process reward models (PRMs), which require expensive human- or LLM-generated labels and often degrade under distribution shifts. In this paper, we introduce AIRL-S, the first natural unification of RL-based and search-based TTS. Central to AIRL-S is the insight that the reward function learned during RL training inherently represents the ideal PRM for guiding downstream search. Specifically, we leverage adversarial inverse reinforcement learning (AIRL) combined with group relative policy optimization (GRPO) to learn a dense, dynamic PRM directly from correct reasoning traces, entirely eliminating the need for labeled intermediate process data. At inference, the resulting PRM simultaneously serves as the critic for RL rollouts and as a heuristic to effectively guide search procedures, facilitating robust reasoning chain extension, mitigating reward hacking, and enhancing cross-task generalization. Experimental results across eight benchmarks, including mathematics, scientific reasoning, and code generation, demonstrate that our unified approach improves performance by 9 % on average over the base model, matching GPT-4o. Furthermore, when integrated into multiple search algorithms, our PRM consistently outperforms all baseline PRMs trained with labeled data. These results underscore that, indeed, your reward function for RL is your best PRM for search, providing a robust and cost-effective solution to complex reasoning tasks in LLMs.
Authors: Viktor Valadi, Mattias {\AA}kesson, Johan \"Ostman, Fazeleh Hoseini, Salman Toor, Andreas Hellander
Abstract: Gradient inversion attacks are often presented as a serious privacy threat in federated learning, with recent work reporting increasingly strong reconstructions under favorable experimental settings. However, it remains unclear whether such attacks are feasible in modern, performance-optimized systems deployed in practice. In this work, we evaluate the practical feasibility of gradient inversion for image-based federated learning. We conduct a systematic study across multiple datasets and tasks, including image classification and object detection, using canonical vision architectures at contemporary resolutions. Our results show that while gradient inversion remains possible for certain legacy or transitional designs under highly restrictive assumptions, modern, performance-optimized models consistently resist meaningful reconstruction visually. We further demonstrate that many reported successes rely on upper-bound settings, such as inference mode operation or architectural simplifications which do not reflect realistic training pipelines. Taken together, our findings indicate that, under an honest-but-curious server assumption, high-fidelity image reconstruction via gradient inversion does not constitute a critical privacy risk in production-optimized federated learning systems, and that practical risk assessments must carefully distinguish diagnostic attack settings from real-world deployments.
Authors: Liana Patel, Negar Arabzadeh, Harshit Gupta, Ankita Sundar, Ion Stoica, Matei Zaharia, Carlos Guestrin
Abstract: The ability to research and synthesize knowledge is central to human expertise and progress. A new class of AI systems--designed for generative research synthesis--aims to automate this process by retrieving information from the live web and producing long-form, cited reports. Yet, evaluating such systems remains an open challenge: existing question-answering benchmarks focus on short, factual answers, while expert-curated datasets risk staleness and data contamination. Neither captures the complexity and evolving nature of real research synthesis tasks. We introduce DeepScholar-bench, a live benchmark and automated evaluation framework for generative research synthesis. DeepScholar-bench draws queries and human-written exemplars from recent, high-quality ArXiv papers and evaluates a real synthesis task: generating a related work section by retrieving, synthesizing, and citing prior work. Our automated framework holistically measures performance across three key dimensions--knowledge synthesis, retrieval quality, and verifiability. To further future work, we also contribute DeepScholar-ref, a simple, open-source reference pipeline, which is implemented on the LOTUS framework and provides a strong baseline. Using DeepScholar-bench, we systematically evaluate prior open-source systems, search agents with strong models, OpenAI's DeepResearch, and DeepScholar-ref. We find DeepScholar-bench is far from saturated: no system surpasses a geometric mean of $31\%$ across all metrics. These results highlight both the difficulty and importance of DeepScholar-bench as a foundation for advancing AI systems capable of generative research synthesis. We make our benchmark code and data available at https://github.com/guestrin-lab/deepscholar-bench.
Authors: Hanzhen Wang, Jiaming Xu, Yushun Xiang, Jiayi Pan, Yongkang Zhou, Yong-Lu Li, Guohao Dai
Abstract: Pruning is a typical acceleration technique for compute-bound models by removing computation on unimportant values. Recently, it has been applied to accelerate Vision-Language-Action (VLA) model inference. However, existing acceleration methods focus on local information from the current action step and ignore the global context, leading to >20% success rate drop and limited speedup in some scenarios. In this paper, we point out spatial-temporal consistency in VLA tasks: input images in consecutive steps exhibit high similarity, and propose the key insight that token selection should combine local information with global context of the model. Based on this, we propose SpecPrune-VLA, a training-free, two-level pruning method with heuristic control. (1) Action-level static pruning. We leverage global history and local attention to statically reduce visual tokens per action. (2) Layer-level dynamic pruning. We prune tokens adaptively per layer based on layer-wise importance. (3) Lightweight action-aware controller: We classify actions as coarse- or fine-grained by the speed of the end effector and adjust pruning aggressiveness accordingly. Extensive experiments show that SpecPrune-VLA achieves up to 1.57$\times$ speedup in LIBERO simulation and 1.70$\times$ on real-world tasks, with negligible success rate degradation.
Authors: Iv\'an Vicente Moreno Cencerrado, Arnau Padr\'es Masdemont, Anton Gonzalvez Hawthorne, David Demitri Africa, Lorenzo Pacchiardi
Abstract: Do large language models (LLMs) anticipate when they will answer correctly? To study this, we extract activations after a question is read but before any tokens are generated, and train linear probes to predict whether the model's forthcoming answer will be correct. Across three open-source model families ranging from 7 to 70 billion parameters, projections on this "in-advance correctness direction" trained on generic trivia questions predict success in distribution and on diverse out-of-distribution knowledge datasets, indicating a deeper signal than dataset-specific spurious features, and outperforming black-box baselines and verbalised predicted confidence. Predictive power saturates in intermediate layers and, notably, generalisation falters on questions requiring mathematical reasoning. Moreover, for models responding "I don't know", doing so strongly correlates with the probe score, indicating that the same direction also captures confidence. By complementing previous results on truthfulness and other behaviours obtained with probes and sparse auto-encoders, our work contributes essential findings to elucidate LLM internals.
Authors: Jie Li, Youyang Hou, Laura Lin, Ruihao Zhu, Hancheng Cao, Abdallah El Ali
Abstract: Generative AI is reshaping product design practices through "vibe coding", where product team members express intent in natural language and AI translates it into functional prototypes and code. Despite rapid adoption, little research has examined how vibe coding reconfigures product development workflows and collaboration. Drawing on interviews with 22 product team members across enterprises, startups, and academia, we show how vibe coding follows a four-stage workflow of ideation, generation, debugging, and review. This accelerates iteration, supports creativity, and lowers participation barriers. However, participants reported challenges of code unreliability, integration, and AI over-reliance. We find tensions between efficiency-driven prototyping ("intending the right design") and reflection ("designing the right intention"), introducing new asymmetries in trust, responsibility, and social stigma within teams. Through a responsible human-AI collaboration lens for AI-assisted product design and development, we contribute a deeper understanding of deskilling, ownership and disclosure, and creativity safeguarding in the age of vibe coding.
Authors: Qi Wang, Shituo Ma, Guoxin Yu, Hanyang Peng, Yue Yu
Abstract: Voice cloning for Text-to-Speech (TTS) aims to generate expressive and personalized speech from text using limited data from a target speaker. Federated Learning (FL) offers a collaborative and privacy-preserving framework for this task, but existing approaches suffer from high communication costs and tend to suppress stylistic heterogeneity, resulting in insufficient personalization. To address these issues, we propose Fed-PISA, which stands for Federated Personalized Identity-Style Adaptation. To minimize communication costs, Fed-PISA introduces a disentangled Low-Rank Adaptation (LoRA) mechanism: the speaker's timbre is retained locally through a private ID-LoRA, while only a lightweight style-LoRA is transmitted to the server, thereby minimizing parameter exchange. To harness heterogeneity, our aggregation method, inspired by collaborative filtering, is introduced to create custom models for each client by learning from stylistically similar peers. Experiments show that Fed-PISA improves style expressivity, naturalness, and speaker similarity, outperforming standard federated baselines with minimal communication costs.
Authors: Xinyu Liu, Guolei Sun, Cheng Wang, Yixuan Yuan, Ender Konukoglu
Abstract: High-resolution (HR) medical videos are vital for accurate diagnosis, yet are hard to acquire due to hardware limitations and physiological constraints. Clinically, the collected low-resolution (LR) medical videos present unique challenges for video super-resolution (VSR) models, including camera shake, noise, and abrupt frame transitions, which result in significant optical flow errors and alignment difficulties. Additionally, tissues and organs exhibit continuous and nuanced structures, but current VSR models are prone to introducing artifacts and distorted features that can mislead doctors. To this end, we propose MedVSR, a tailored framework for medical VSR. It first employs Cross State-Space Propagation (CSSP) to address the imprecise alignment by projecting distant frames as control matrices within state-space models, enabling the selective propagation of consistent and informative features to neighboring frames for effective alignment. Moreover, we design an Inner State-Space Reconstruction (ISSR) module that enhances tissue structures and reduces artifacts with joint long-range spatial feature learning and large-kernel short-range information aggregation. Experiments across four datasets in diverse medical scenarios, including endoscopy and cataract surgeries, show that MedVSR significantly outperforms existing VSR models in reconstruction performance and efficiency. Code released at https://github.com/CUHK-AIM-Group/MedVSR.
Authors: Lihui Liu, Jiayuan Ding, Subhabrata Mukherjee, Carl J. Yang
Abstract: Large Language Models (LLMs) have achieved impressive performance across a wide range of applications. However, they often suffer from hallucinations in knowledge-intensive domains due to their reliance on static pretraining corpora. To address this limitation, Retrieval-Augmented Generation (RAG) enhances LLMs by incorporating external knowledge sources during inference. Among these sources, textual graphs provide structured and semantically rich information that supports more precise and interpretable reasoning. This has led to growing interest in graph-based RAG systems. Despite their potential, most existing approaches rely on a single retriever to identify relevant subgraphs, which limits their ability to capture the diverse aspects of complex queries. Moreover, these systems often struggle to accurately judge the relevance of retrieved content, making them prone to distraction by irrelevant noise. To address these challenges, in this paper, we propose MIXRAG, a Mixture-of-Experts Graph-RAG framework that introduces multiple specialized graph retrievers and a dynamic routing controller to better handle diverse query intents. Each retriever is trained to focus on a specific aspect of graph semantics, such as entities, relations, or subgraph topology. A Mixture-of-Experts module adaptively selects and fuses relevant retrievers based on the input query. To reduce noise in the retrieved information, we introduce a query-aware GraphEncoder that carefully analyzes relationships within the retrieved subgraphs, highlighting the most relevant parts while down-weighting unnecessary noise. Empirical results demonstrate that our method achieves state-of-the-art performance and consistently outperforms various baselines. MIXRAG is effective across a wide range of graph-based tasks in different domains. The code will be released upon paper acceptance.
Authors: Thanh-Long V. Le, Myeongho Jeon, Kim Vu, Viet Lai, Eunho Yang
Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) is a powerful framework for improving the reasoning abilities of Large Language Models (LLMs). However, current methods such as GRPO rely only on problems where the model responses to the same input differ in correctness, while ignoring those where all responses receive the same reward -- so-called zero-variance prompts. In this work, we argue that such prompts are not useless but can, in fact, provide meaningful feedback for policy optimization. To this end, we introduce Reinforcement Learning with Zero-Variance Prompts (RL-ZVP), a novel algorithm that extract learning signals from zero-variance prompts. RL-ZVP directly rewards correctness and penalizes errors even without contrasting responses, modulating feedback with token-level characteristics to preserve informative, nuanced signals. Across six math reasoning benchmarks, RL-ZVP achieves significant improvements of up to 8.61 points in accuracy and 7.77 points in pass rate over GRPO, while consistently outperforming other baselines that filter out zero-variance prompts. These results highlight the untapped potential of learning from zero-variance prompts in RLVR. The project page is available at https://bltnynk.github.io/publications/rl-zvp/.
Authors: Pavan Karjol, Vivek V Kashyap, Rohan Kashyap, Prathosh A P
Abstract: We propose a modular, data-driven framework for jointly learning unknown functional mappings and discovering the underlying one-parameter symmetry subgroup governing the data. Unlike conventional geometric deep learning methods that assume known symmetries, our approach identifies the relevant continuous subgroup directly from data. We consider the broad class of one-parameter subgroups, which admit a canonical geometric classification into three regimes: elliptical, hyperbolic, and parabolic. Given an assumed regime, our framework instantiates a corresponding symmetry discovery architecture with invariant and equivariant representation layers structured according to the Lie algebra of the subgroup, and learns the exact generator parameters end-to-end from data. This yields models whose invariance or equivariance is guaranteed by construction and admits formal proofs, enabling symmetry to be explicitly traced to identifiable components of the architecture. The approach is applicable to one-parameter subgroups of a wide range of matrix Lie groups, including $SO(n)$, $SL(n)$, and the Lorentz group. Experiments on synthetic and real-world systems, including moment of inertia prediction, double-pendulum dynamics, and high-energy \textit{Top Quark Tagging}, demonstrate accurate subgroup recovery and strong predictive performance across both compact and non-compact regimes.
Authors: Xiaoyang Liu, Tao Zhu, Zineng Dong, Yuntian Liu, Qingfeng Guo, Zhaoxuan Liu, Yu Chen, Tao Luo
Abstract: Despite significant strides in statement autoformalization, a critical gap remains in the development of automated evaluation metrics capable of assessing formal translation quality. Existing metrics often fail to balance semantic and structural information: string-based methods neglect semantics, whereas proof-based approaches offer no graded similarity when proofs fail. To address these issues, we introduce ASSESS (A Semantic and Structural Evaluation Framework for Statement Similarity), which captures syntactic structure by transforming formal statements into operator trees and computes a real-valued similarity score using our novel TransTED (Transformation Tree Edit Distance) Similarity metric by incorporating semantic transformations. For rigorous validation, we present EPLA (Evaluating Provability and Likeness for Autoformalization), a benchmark comprising 1,247 expert-annotated formal statement pairs derived from miniF2F and ProofNet, distinctively labeled for both semantic provability and structural likeness. Experiments on the EPLA benchmark demonstrate that TransTED Similarity surpasses existing methods, achieving state-of-the-art accuracy and Kappa score. The benchmark dataset, code, and detailed experimental results are available at https://github.com/XiaoyangLiu-sjtu/ASSESS.
Authors: Qinxun Bai, Yuxuan Han, Wei Xu, Zhengyuan Zhou
Abstract: The actor-critic (AC) framework has achieved strong empirical success in off-policy reinforcement learning but suffers from the "moving target" problem, where the evaluated policy changes continually. Functional critics, or policy-conditioned value functions, address this by explicitly including a representation of the policy as input. While conceptually appealing, previous efforts have struggled to remain competitive against standard AC. In this work, we revisit functional critics within the actor-critic framework and identify two critical aspects that render them a necessity rather than a luxury. First, we demonstrate their power in stabilizing the complex interplay between the "deadly triad" and the "moving target". We provide a convergent off-policy AC algorithm under linear functional approximation that dismantles several longstanding barriers between theory and practice: it utilizes target-based TD learning, accommodates dynamic behavior policies, and operates without the restrictive "full coverage" assumptions. By formalizing a dual trust-coverage mechanism, our framework provides principled guidelines for pursuing sample efficiency-rigorously governing behavior policy updates and critic re-evaluations to maximize off-policy data utility. Second, we uncover a foundational link between functional critics and efficient exploration. We demonstrate that existing model-free approximations of posterior sampling are limited in capturing policy-dependent uncertainty, a gap the functional critic formalism bridges. These results represent, to our knowledge, first-of-their-kind contributions to the RL literature. Practically, we propose a tailored neural network architecture and a minimalist AC algorithm. In preliminary experiments on the DeepMind Control Suite, this implementation achieves performance competitive with state-of-the-art methods without standard implementation heuristics.
Authors: Yuqiao Meng, Luoxi Tang, Feiyang Yu, Jinyuan Jia, Guanhua Yan, Ping Yang, Zhaohan Xi
Abstract: Large language models (LLMs) are increasingly used to help security analysts manage the surge of cyber threats, automating tasks from vulnerability assessment to incident response. Yet in operational CTI workflows, reliability gaps remain substantial. Existing explanations often point to generic model issues (e.g., hallucination), but we argue the dominant bottleneck is the threat landscape itself: CTI is heterogeneous, volatile, and fragmented. Under these conditions, evidence is intertwined, crowdsourced, and temporally unstable, which are properties that standard LLM-based studies rarely capture. In this paper, we present a comprehensive empirical study of LLM vulnerabilities in CTI reasoning. We introduce a human-in-the-loop categorization framework that robustly labels failure modes across the CTI lifecycle, avoiding the brittleness of automated "LLM-as-a-judge" pipelines. We identify three domain-specific cognitive failures: spurious correlations from superficial metadata, contradictory knowledge from conflicting sources, and constrained generalization to emerging threats. We validate these mechanisms via causal interventions and show that targeted defenses reduce failure rates significantly. Together, these results offer a concrete roadmap for building resilient, domain-aware CTI agents.
Authors: Xin Qiu, Yulu Gan, Conor F. Hayes, Qiyao Liang, Yinggan Xu, Roberto Dailey, Elliot Meyerson, Babak Hodjat, Risto Miikkulainen
Abstract: Fine-tuning large language models (LLMs) for downstream tasks is an essential stage of modern AI deployment. Reinforcement learning (RL) has emerged as the dominant fine-tuning paradigm, underpinning many state-of-the-art LLMs. In contrast, evolution strategies (ES) has largely been overlooked due to the widespread belief that it does not scale to modern model sizes. This paper overturns this assumption by demonstrating the first successful application of ES to full-parameter fine-tuning of LLMs at the billion-parameter scale, without dimensionality reduction. ES can indeed search over extremely high-dimensional parameter spaces and outperform established RL implementations across multiple axes, including improved tolerance to long-horizon and delayed rewards, robustness across diverse base LLMs, reduced susceptibility to reward hacking, and improved training stability. These findings suggest that ES is not merely a viable alternative to RL, but a fundamentally different and powerful backpropagation-free post-training paradigm that opens a new direction for LLM fine-tuning beyond current RL-based approaches. The source codes are provided at: https://github.com/VsonicV/es-fine-tuning-paper.
Authors: Haijier Chen, Bo Xu, Shoujian Zhang, Haoze Liu, Jiaxuan Lin, Jingrong Wang
Abstract: Recent developments in Multimodal Large Language Models (MLLMs) have significantly improved Vision-Language (VL) reasoning in 2D domains. However, extending these capabilities to 3D scene understanding remains a major challenge. Existing 3D Multimodal Large Language Models (3D-MLLMs) often depend on 3D data inputs, which limits scalability and generalization. To address this limitation, we propose Vid-LLM, a video-based 3D-MLLM that directly processes video inputs without requiring external 3D data, making it practical for real-world deployment. In our method, the geometric prior are directly used to improve the performance of the sceen perception. To integrate the geometric cues into the MLLM compactly, we design a Cross-Task Adapter (CTA) module to align the 3D geometric priors with the vision-language representations. To ensure geometric consistency and integrity, we introduce a Metric Depth Model that recovers real-scale geometry from the reconstruction outputs. Finally, the model is fine-tuned with a two-stage distillation optimization strategy, realizing fast convergence and stabilizes training. Extensive experiments across diverse benchmarks verified the effectiveness of our method on 3D Question Answering, 3D Dense Captioning and 3D Visual Grounding tasks, demonstrating the superior multi-task capabilities.
Authors: Jonas H\"ubotter, Patrik Wolf, Alexander Shevchenko, Dennis J\"uni, Andreas Krause, Gil Kur
Abstract: Recent empirical studies have explored the idea of continuing to train a model at test-time for a given task, known as test-time training (TTT), and have found it to yield significant performance improvements. However, there is limited understanding of why and when TTT is effective. Earlier explanations mostly focused on the observation that TTT may help when applied to out-of-distribution adaptation or used with privileged data. However, the growing scale of foundation models with most test data being in-distribution questions these explanations. We instead posit that foundation models remain globally underparameterized, with TTT providing a mechanism for specialization after generalization, focusing capacity on concepts relevant to the test task. Specifically, under the linear representation hypothesis, we propose a model in which TTT achieves a substantially smaller in-distribution test error than global training. We empirically validate our model's key assumptions by training a sparse autoencoder on ImageNet, showing that semantically related data points are explained by only a few shared concepts. Finally, we perform scaling studies across image and language tasks that confirm the practical implications of our model, identifying the regimes where specialization is most effective.
Authors: Jiacheng Shi, Hongfei Du, Yangfan He, Y. Alicia Hong, Ye Gao
Abstract: Emotional text-to-speech seeks to convey affect while preserving intelligibility and prosody, yet existing methods rely on coarse labels or proxy classifiers and receive only utterance-level feedback. We introduce Emotion-Aware Stepwise Preference Optimization (EASPO), a post-training framework that aligns diffusion TTS with fine-grained emotional preferences at intermediate denoising steps. Central to our approach is EASPM, a time-conditioned model that scores noisy intermediate speech states and enables automatic preference pair construction. EASPO optimizes generation to match these stepwise preferences, enabling controllable emotional shaping. Experiments show superior performance over existing methods in both expressiveness and naturalness.
Authors: Luis Burbano, Diego Ortiz, Qi Sun, Siwei Yang, Haoqin Tu, Cihang Xie, Yinzhi Cao, Alvaro A Cardenas
Abstract: Embodied Artificial Intelligence (AI) promises to handle edge cases in robotic vehicle systems where data is scarce by using common-sense reasoning grounded in perception and action to generalize beyond training distributions and adapt to novel real-world situations. These capabilities, however, also create new security risks. In this paper, we introduce CHAI (Command Hijacking against embodied AI), a physical environment indirect prompt injection attack that exploits the multimodal language interpretation abilities of AI models. CHAI embeds deceptive natural language instructions, such as misleading signs, in visual input, systematically searches the token space, builds a dictionary of prompts, and guides an attacker model to generate Visual Attack Prompts. We evaluate CHAI on four LVLM agents: drone emergency landing, autonomous driving, aerial object tracking, and on a real robotic vehicle. Our experiments show that CHAI consistently outperforms state-of-the-art attacks. By exploiting the semantic and multimodal reasoning strengths of next-generation embodied AI systems, CHAI underscores the urgent need for defenses that extend beyond traditional adversarial robustness.
Authors: Yongchao Long, Xian Wu, Yingying Zhang, Xianbin Wen, Yuxi Zhou, Shenda Hong
Abstract: While Retrieval-Augmented Generation (RAG) enables large language models (LLMs) to generate contextually grounded responses, contextual faithfulness remains challenging as LLMs may not consistently trust provided context, leading to hallucinations that undermine reliability. We observe an inverse correlation between response copying degree and context-unfaithful hallucinations on RAGTruth, suggesting that higher copying degrees reduce hallucinations by fostering genuine contextual belief. We propose CopyPasteLLM, obtained through two-stage high-copying response preference training. We design three prompting methods to enhance copying degree, demonstrating that high-copying responses achieve superior contextual faithfulness and hallucination control. These approaches enable a fully automated pipeline that transforms generated responses into high-copying preference data for training CopyPasteLLM. On FaithEval, ConFiQA and PubMedQA, CopyPasteLLM achieves best performance in both counterfactual and original contexts, remarkably with 12.2% to 24.5% accuracy improvements on FaithEval over the best baseline, while requiring only 365 training samples -- 1/50th of baseline data. To elucidate CopyPasteLLM's effectiveness, we propose the Context-Parameter Copying Capturing algorithm. Interestingly, this reveals that CopyPasteLLM recalibrates reliance on internal parametric knowledge rather than external knowledge during generation. All codes are available at https://github.com/longyongchao/CopyPasteLLM
Authors: Bonaventure F. P. Dossou, Henri A\"idasso
Abstract: Natural Language Processing (NLP) for low-resource languages remains fundamentally constrained by the lack of textual corpora, standardized orthographies, and scalable annotation pipelines. While recent advances in large language models have improved cross-lingual transfer, they remain inaccessible to underrepresented communities due to their reliance on massive, pre-collected data and centralized infrastructure. In this position paper, we argue for a paradigm shift toward open-ended, interactive language discovery, where AI systems learn new languages dynamically through dialogue rather than static datasets. We contend that the future of language technology, particularly for low-resource and under-documented languages, must move beyond static data collection pipelines toward interactive, uncertainty-driven discovery, where learning emerges dynamically from human-machine collaboration instead of being limited to pre-existing datasets. We propose a framework grounded in joint human-machine uncertainty, combining epistemic uncertainty from the model with hesitation cues and confidence signals from human speakers to guide interaction, query selection, and memory retention. This paper is a call to action: we advocate a rethinking of how AI engages with human knowledge in under-documented languages, moving from extractive data collection toward participatory, co-adaptive learning processes that respect and empower communities while discovering and preserving the world's linguistic diversity. This vision aligns with principles of human-centered AI, emphasizing interactive, cooperative model building between AI systems and speakers.
Authors: Zifan Song, Kaitao Song, Guosheng Hu, Ding Qi, Junyao Gao, Xiaohua Wang, Dongsheng Li, Cairong Zhao
Abstract: Recent advancements in large language models (LLMs) and agentic systems have shown exceptional decision-making capabilities, revealing significant potential for autonomic finance. Current financial trading agents predominantly simulate anthropomorphic roles that inadvertently introduce emotional biases and rely on peripheral information, while being constrained by the necessity for continuous inference during deployment. In this paper, we pioneer the harmonization of strategic depth in agents with the mechanical rationality essential for quantitative trading. Consequently, we present TiMi (Trade in Minutes), a rationality-driven multi-agent system that architecturally decouples strategy development from minute-level deployment. TiMi leverages specialized LLM capabilities of semantic analysis, code programming, and mathematical reasoning within a comprehensive policy-optimization-deployment chain. Specifically, we propose a two-tier analytical paradigm from macro patterns to micro customization, layered programming design for trading bot implementation, and closed-loop optimization driven by mathematical reflection. Extensive evaluations across 200+ trading pairs in stock and cryptocurrency markets empirically validate the efficacy of TiMi in stable profitability, action efficiency, and risk control under volatile market dynamics.
Authors: Ye Hong, Yatao Zhang, Konrad Schindler, Martin Raubal
Abstract: Understanding and modeling human mobility is central to challenges in transport planning, sustainable urban design, and public health. Despite decades of effort, simulating individual mobility remains challenging because of its complex, context-dependent, and exploratory nature. Here, we advance a unified event-level formulation of daily mobility and propose MobilityGen to generate multi-attribute event sequences over days to weeks at large spatial scales. By linking behavioral attributes with environmental context, MobilityGen reproduces key patterns such as scaling laws for location visits, activity time allocation, and the coupled evolution of travel mode and destination choices. It reflects spatio-temporal variability and generates diverse and plausible mobility patterns consistent with the built environment. Beyond standard validation, MobilityGen enables analyses that have been difficult with earlier models, including how access to urban space varies across travel modes and how co-presence dynamics shape social exposure and segregation. Together, these results support an integrated, data-driven basis for fine-grained studies of human mobility behavior and its societal implications.
Authors: Yiqin Lv, Zhiyu Mou, Miao Xu, Jinghao Chen, Qi Wang, Yixiu Mao, Yun Qu, Rongquan Bai, Chuan Yu, Jian Xu, Bo Zheng, Xiangyang Ji
Abstract: Generative auto-bidding has demonstrated strong performance in online advertising, yet it often suffers from data scarcity in small-scale settings with limited advertiser participation. While cross-task data sharing is a natural remedy to mitigate this issue, naive approaches often introduce gradient bias due to distribution shifts across different tasks, and existing methods are not readily applicable to generative auto-bidding. In this paper, we propose Validation-Aligned Optimization (VAO), a principled data-sharing method that adaptively reweights cross-task data contributions based on validation performance feedback. Notably, VAO aligns training dynamics to prioritize updates that improve generalization on the target task, effectively leveraging auxiliary data and mitigating gradient bias. Building on VAO, we introduce a unified generative autobidding framework that generalizes across multiple tasks using a single model and all available task data. Extensive experiments on standard auto-bidding benchmarks validate the effectiveness of our approach.
Authors: Xiangyuan Xue, Yifan Zhou, Guibin Zhang, Zaibin Zhang, Yijiang Li, Chen Zhang, Zhenfei Yin, Philip Torr, Wanli Ouyang, Lei Bai
Abstract: Self-evolution is a central research topic in enabling large language model (LLM)-based agents to continually improve their capabilities after pretraining. Recent research has witnessed a transition from reinforcement learning (RL)-free to RL-based methods. Current RL-based methods either rely on dense external reward signals or extract intrinsic reward signals from LLMs themselves. However, these approaches diverge from the self-evolution mechanisms observed in human intelligence, where individuals learn and improve through mutual discussion and collaboration. In this work, we introduce Co-Evolving Multi-Agent Systems (CoMAS), a novel framework that enables agents to improve autonomously by learning from inter-agent interactions without external supervision. CoMAS generates intrinsic rewards from rich discussion dynamics, employs an LLM-as-a-judge mechanism to formulate these rewards, and optimizes each agent's policy through RL, thereby enabling decentralized and scalable co-evolution. Experimental results demonstrate that CoMAS consistently outperforms untrained agents and achieves state-of-the-art performance across most evaluation settings. Ablation studies confirm the necessity of interaction-based reward signals and reveal promising scalability as the number and diversity of agents increase. These findings establish CoMAS as a novel and effective paradigm for self-evolution in LLM-based agents.
Authors: Nan Lu, Yurong Hu, Jiaquan Fang, Yan Liu, Rui Dong, Yiming Wang, Rui Lin, Shaoyi Xu
Abstract: Effective e-commerce risk management requires in-depth case investigations to identify emerging fraud patterns in highly adversarial environments. However, manual investigation typically requires analyzing the associations and couplings among multi-source heterogeneous data, a labor-intensive process that limits efficiency. While Large Language Models (LLMs) show promise in automating these analyses, their deployment is hindered by the complexity of risk scenarios and the sparsity of long-tail domain knowledge. To address these challenges, we propose Sherlock, a framework that integrates structured domain knowledge with LLM-based reasoning through three core modules. First, we construct a domain Knowledge Base (KB) by distilling structured expertise from heterogeneous knowledge sources. Second, we design a two-stage retrieval-augmented generation strategy tailored for case investigation, which combines input contextual augmentation with a Reflect & Refine module to fully leverage the KB for improved analysis quality. Finally, we develop an integrated platform for operations and annotation to drive a self-evolving data flywheel. By combining real-time hotfixes through KB updates with periodic logic alignment via post-training, we facilitate continuous system evolution to counteract adversarial drifts. Online A/B tests at JD dot com demonstrate that Sherlock achieves an 82% Expert Acceptance Rate (EAR) and a 386.7% increase in daily investigation throughput. An additional 90-day evaluation shows that the flywheel successfully recovers from performance decay caused by changing tactics twice, raising the EAR ceiling by around 3.5% through autonomous model updates.
Authors: Xu Pan, Ely Hahami, Jingxuan Fan, Ziqian Xie, Haim Sompolinsky
Abstract: Large language models (LLMs) are often used in environments where facts evolve, yet factual knowledge updates via fine-tuning on unstructured text often suffers from 1) reliance on compute-heavy paraphrase augmentation and 2) the reversal curse. Recent studies show diffusion large language models (dLLMs) require fewer training samples to achieve lower loss in pre-training and are more resistant to the reversal curse, suggesting dLLMs may learn new knowledge more easily than autoregressive LLMs (arLLMs). We test this hypothesis in controlled knowledge fine-tuning experiments and find that while arLLMs rely on paraphrase augmentation to generalize knowledge text into question-answering (QA) capability, dLLMs do not require paraphrases to achieve high QA accuracy. To further investigate whether the demasking objective alone can induce such a knowledge injection advantage in dLLMs regardless of their diffusion denoising paradigm, we propose masked fine-tuning for arLLMs, which prompts an arLLM to reconstruct the original text given a masked version in context. The masked fine-tuning for arLLMs substantially improves the efficacy of knowledge injection, i.e. no paraphrase needed and resistant to the reversal curse, closing the gap between arLLMs and dLLMs. We also demonstrate that the same demasking objective improves supervised fine-tuning (SFT) on math tasks over standard SFT, suggesting broader applicability of the demasking objective.
Authors: Parsa Gooya, Reinel Sospedra-Alfonso
Abstract: Seasonal forecast of Arctic sea ice concentration is key to mitigate the negative impact and assess potential opportunities posed by the rapid decline of sea ice coverage. Seasonal prediction systems based on climate models often show systematic biases and complex spatio-temporal errors that grow with the forecasts. Consequently, operational predictions are routinely bias corrected and calibrated using retrospective forecasts. For predictions of Arctic sea ice concentration, error corrections are mainly based on one-to-one post-processing methods including climatological mean or linear regression correction and, more recently, machine learning. Such deterministic adjustments are confined at best to the limited number of costly-to-run ensemble members of the raw forecast. However, decision-making requires proper quantification of uncertainty and likelihood of events, particularly of extremes. We introduce a probabilistic error correction framework based on a conditional Variational Autoencoder model to map the conditional distribution of observations given the biased model prediction. This method naturally allows for generating large ensembles of adjusted forecasts. We evaluate our model using deterministic and probabilistic metrics and show that the adjusted forecasts are better calibrated, closer to the observational distribution, and have smaller errors than climatological mean adjusted forecasts.
Authors: Abdelkrim Zitouni, Mehdi Hennequin, Juba Agoun, Ryan Horache, Nadia Kabachi, Omar Rivasplata
Abstract: We derive a novel PAC-Bayesian generalization bound for reinforcement learning that explicitly accounts for Markov dependencies in the data, through the chain's mixing time. This contributes to overcoming challenges in obtaining generalization guarantees for reinforcement learning, where the sequential nature of data breaks the independence assumptions underlying classical bounds. The new bound provides non-vacuous certificates for modern off-policy algorithms such as Soft Actor-Critic. We demonstrate the practical utility of the bound through PB-SAC, a novel algorithm that optimizes the bound during training to guide exploration. Experiments across several continuous control tasks show that the proposed approach provides meaningful confidence certificates while maintaining competitive performance.
Authors: Yubo Zhang, Xiao-Yang Liu, Xiaodong Wang
Abstract: We develop an unsupervised deep learning framework for downlink beamforming in large-scale MU-MISO channels. The model is trained offline, allowing real-time inference through lightweight feedforward computations in dynamic communication environments. Following the learning-to-optimize (L2O) paradigm, a multi-layer Transformer iteratively refines both channel and beamformer features via residual connections. To enhance training, three strategies are introduced: (i) curriculum learning (CL) to improve early-stage convergence and avoid local optima, (ii) semi-amortized learning to refine each Transformer block with a few gradient ascent steps, and (iii) sliding-window training to stabilize optimization by training only a subset of Transformer blocks at a time. Extensive simulations show that the proposed scheme outperforms existing baselines at low-to-medium SNRs and closely approaches WMMSE performance at high SNRs, while achieving substantially faster inference than iterative and online learning approaches.
Authors: Jing Yang, Qiyao Wei, Jiaxin Pei
Abstract: The rapid growth of AI conferences is straining an already fragile peer-review system, leading to heavy reviewer workloads, expertise mismatches, inconsistent evaluation standards, superficial or templated reviews, and limited accountability under compressed timelines. In response, conference organizers have introduced new policies and interventions to preserve review standards. Yet these ad-hoc changes often create further concerns and confusion about the review process, leaving how papers are ultimately accepted - and how practices evolve across years - largely opaque. We present Paper Copilot, a system that creates durable digital archives of peer reviews across a wide range of computer-science venues, an open dataset that enables researchers to study peer review at scale, and a large-scale empirical analysis of ICLR reviews spanning multiple years. By releasing both the infrastructure and the dataset, Paper Copilot supports reproducible research on the evolution of peer review. We hope these resources help the community track changes, diagnose failure modes, and inform evidence-based improvements toward a more robust, transparent, and reliable peer-review system.
Authors: Filipe Laitenberger, Dawid Kopiczko, Cees G. M. Snoek, Yuki M. Asano
Abstract: We introduce GateSkip, a simple residual-stream gating mechanism that enables token-wise layer skipping in decoder-only LMs. Each Attention/MLP branch is equipped with a sigmoid-linear gate that condenses the branch's output before it re-enters the residual stream. During inference we rank tokens by the gate values and skip low-importance ones using a per-layer budget. While early-exit or router-based Mixture-of-Depths models are known to be unstable and need extensive retraining, our smooth, differentiable gates fine-tune stably on top of pretrained models. On long-form reasoning, we save up to 15% compute while retaining over 90% of baseline accuracy. For increasingly larger models, this tradeoff improves drastically. On instruction-tuned models we see accuracy gains at full compute and match baseline quality near 50% savings. The learned gates give insight into transformer information flow (e.g., BOS tokens act as anchors), and the method combines easily with quantization, pruning, and self-speculative decoding.
Authors: Lukas Helff, Ruben H\"arle, Wolfgang Stammer, Felix Friedrich, Manuel Brack, Antonia W\"ust, Hikaru Shindo, Patrick Schramowski, Kristian Kersting
Abstract: Large language models (LLMs) excel at generating fluent text, but their internal reasoning remains opaque and difficult to control. Sparse autoencoders (SAEs) make hidden activations more interpretable by exposing latent features that often align with human concepts. Yet, these features are fragile and passive, offering no mechanism for systematic reasoning or model control. To address this, we introduce ActivationReasoning (AR), a framework that embeds explicit logical reasoning into the latent space of LLMs. It proceeds in three stages: (1) Finding latent representations, first latent concept representations are identified (e.g., via SAEs) and organized into a dictionary; (2) Activating propositions, at inference time AR detects activating concepts and maps them to logical propositions; and (3)Logical reasoning, applying logical rules over these propositions to infer higher-order structures, compose new concepts, and steer model behavior. We evaluate AR on multi-hop reasoning (PrOntoQA), abstraction and robustness to indirect concept cues (Rail2Country), reasoning over natural and diverse language (ProverQA), and context-sensitive safety (BeaverTails). Across all tasks, AR scales robustly with reasoning complexity, generalizes to abstract and context-sensitive tasks, and transfers across model backbones. These results demonstrate that grounding logical structure in latent activations not only improves transparency but also enables structured reasoning, reliable control, and alignment with desired behaviors, providing a path toward more reliable and auditable AI.
Authors: Alan Saji, Raj Dabre, Anoop Kunchukuttan, Ratish Puduppully
Abstract: Large Reasoning Models (LRMs) achieve strong performance on mathematical, scientific, and other question-answering tasks, but their multilingual reasoning abilities remain underexplored. When presented with non-English questions, LRMs often default to reasoning in English, raising concerns about interpretability and the handling of linguistic and cultural nuances. We systematically compare an LRM's reasoning in English versus the language of the question. Our evaluation spans two tasks: MGSM and GPQA Diamond. Beyond measuring answer accuracy, we also analyze cognitive attributes in the reasoning traces. We find that English reasoning traces exhibit a substantially higher presence of these cognitive behaviors, and that reasoning in English generally yields higher final-answer accuracy, with the performance gap increasing as tasks become more complex. However, this English-centric strategy is susceptible to a key failure mode - getting "Lost in Translation," where translation steps lead to errors that would have been avoided by reasoning in the language of the question.
Authors: Anchit Jain, Stephen Bates
Abstract: Decomposing prediction uncertainty into aleatoric (irreducible) and epistemic (reducible) components is critical for the reliable deployment of machine learning systems. While the mutual information between the response variable and model parameters is a principled measure for epistemic uncertainty, it requires access to the parameter posterior, which is computationally challenging to approximate. Consequently, practitioners often rely on probabilistic predictions from deep ensembles to quantify uncertainty, which have demonstrated strong empirical performance. However, a theoretical understanding of their success from a frequentist perspective remains limited. We address this gap by first considering a bootstrap-based estimator for epistemic uncertainty, which we prove is asymptotically correct. Next, we connect deep ensembles to the bootstrap estimator by decomposing it into data variability and training stochasticity; specifically, we show that deep ensembles capture the training stochasticity component. Through empirical studies, we show that this stochasticity component constitutes the majority of epistemic uncertainty, thereby explaining the effectiveness of deep ensembles.
Authors: Jamison Meindl, Yunsheng Tian, Tony Cui, Veronika Thost, Zhang-Wei Hong, Jie Chen, Wojciech Matusik, Mina Konakovi\'c Lukovi\'c
Abstract: Global optimization of expensive, derivative-free black-box functions demands extreme sample efficiency and decision interpretability. While Large Language Models (LLMs) have shown broad capabilities, even state-of-the-art models remain limited in solving continuous black-box optimization tasks and struggle to maintain exploration-exploitation balance. We introduce GPTOpt, an optimization method that equips LLMs with continuous black-box optimization capabilities by fine-tuning Llama 3.1 8B on structured Bayesian optimization (BO) data, including surrogate model information. This provides an explainable framework calibrated to produce surrogate model outputs comparable to a Gaussian process, while keeping the advantages of flexible LLM-based optimization. On a variety of black-box optimization benchmarks, our model shows favorable performance compared to traditional optimizers and transformer-based alternatives, while providing important context and insight into the model's decisions.
Authors: Chandra Vamsi Krishna Alla, Harish Naidu Gaddam, Manohar Kommi
Abstract: Large Language Models (LLMs) face significant computational and memory constraints when processing long contexts, despite growing demand for applications requiring reasoning over extensive documents, multi-session dialogues, and book length texts. While recent advances have extended context windows to 100K-1M tokens, such approaches incur prohibitive costs for resource constrained deployments. We propose BudgetMem, a novel memory augmented architecture that learns what to remember rather than remembering everything. Our system combines selective memory policies with feature based salience scoring (entity density, TF-IDF, discourse markers, position bias) to decide which information merits storage under strict budget constraints. Unlike existing retrieval augmented generation (RAG) systems that store all chunks, BudgetMem employs learned gating mechanisms coupled with BM25 sparse retrieval for efficient information access. Through comprehensive experiments on 700 question answer pairs across short (237 tokens) and long (5K-10K tokens) documents with Llama-3.2-3B-Instruct, we demonstrate that BudgetMem achieves remarkable results on long documents: only 1.0% F1 score degradation while saving 72.4% memory compared to baseline RAG. We validate our approach through budget sensitivity analysis (testing 7 budget ratios), naive baseline comparisons, and document length analysis, showing that BudgetMem's benefits increase with document length. Our work provides a practical pathway for deploying capable long context systems on modest hardware, democratizing access to advanced language understanding capabilities.
Authors: Shiyao Sang
Abstract: This paper challenges a prevailing epistemological assumption in End-to-End Autonomous Driving: that high-performance planning necessitates high-fidelity world reconstruction. Inspired by cognitive science, we propose the Mental Bayesian Causal World Model (MBCWM) and instantiate it as the Tokenized Intent World Model (TIWM), a novel cognitive computing architecture. Its core philosophy posits that intelligence emerges not from pixel-level objective fidelity, but from the Cognitive Consistency between the agent's internal intentional world and physical reality. By synthesizing von Uexk\"ull's $\textit{Umwelt}$ theory, the neural assembly hypothesis, and the triple causal model (integrating symbolic deduction, probabilistic induction, and force dynamics) into an end-to-end embodied planning system, we demonstrate the feasibility of this paradigm on the nuPlan benchmark. Experimental results in open-loop validation confirm that our Belief-Intent Co-Evolution mechanism effectively enhances planning performance. Crucially, in closed-loop simulations, the system exhibits emergent human-like cognitive behaviors, including map affordance understanding, free exploration, and self-recovery strategies. We identify Cognitive Consistency as the core learning mechanism: during long-term training, belief (state understanding) and intent (future prediction) spontaneously form a self-organizing equilibrium through implicit computational replay, achieving semantic alignment between internal representations and physical world affordances. TIWM offers a neuro-symbolic, cognition-first alternative to reconstruction-based planners, establishing a new direction: planning as active understanding, not passive reaction.
Authors: Ankan Mullick, Sukannya Purkayastha, Saransh Sharma, Pawan Goyal, Niloy Ganguly
Abstract: Voice-controlled dialog systems have become immensely popular due to their ability to perform a wide range of actions in response to diverse user queries. These agents possess a predefined set of skills or intents to fulfill specific user tasks. But every system has its own limitations. There are instances where, even for known intents, if any model exhibits low confidence, it results in rejection of utterances that necessitate manual annotation. Additionally, as time progresses, there may be a need to retrain these agents with new intents from the system-rejected queries to carry out additional tasks. Labeling all these emerging intents and rejected utterances over time is impractical, thus calling for an efficient mechanism to reduce annotation costs. In this paper, we introduce IDALC (Intent Detection and Active Learning based Correction), a semi-supervised framework designed to detect user intents and rectify system-rejected utterances while minimizing the need for human annotation. Empirical findings on various benchmark datasets demonstrate that our system surpasses baseline methods, achieving a 5-10% higher accuracy and a 4-8% improvement in macro-F1. Remarkably, we maintain the overall annotation cost at just 6-10% of the unlabelled data available to the system. The overall framework of IDALC is shown in Fig. 1
Authors: Yifei Zhang, Jiashuo Zhang, Mojtaba Safari, Xiaofeng Yang, Liang Zhao
Abstract: Low-dose chest computed tomography (LDCT) inherently captures both pulmonary and cardiac structures, offering a unique opportunity for joint assessment of lung and cardiovascular health. However, most existing approaches treat these domains as independent tasks, overlooking their physiological interplay and shared imaging biomarkers. We propose an Explainable Cross-Disease Reasoning Framework that enables interpretable cardiopulmonary risk assessment from a single LDCT scan. The framework introduces an agentic reasoning process that emulates clinical diagnostic thinking: first perceiving pulmonary findings, then reasoning through established medical knowledge, and finally deriving a cardiovascular judgment with a natural-language rationale. It integrates three components: a Pulmonary Perception Module that summarizes lung abnormalities, an Agentic Pulmonary-to-Cardiac Reasoning Module that infers their cardiovascular implications, and a Cardiac Feature Extractor that encodes structural biomarkers. Their outputs are fused to produce a holistic cardiovascular risk prediction that is both accurate and physiologically grounded. Experiments on the NLST cohort demonstrate that the proposed framework achieves state-of-the-art performance for CVD screening (AUC=0.919) and mortality prediction (AUC=0.838), outperforming single-disease and purely image-based baselines. Beyond quantitative gains, the framework provides human-verifiable reasoning that aligns with cardiological understanding, revealing coherent links between pulmonary abnormalities and cardiac stress mechanisms. Overall, this work establishes a unified and explainable paradigm for cardiovascular analysis from LDCT, bridging the gap between image-based prediction and mechanism-based medical interpretation.
Authors: Belinda Z. Li, Zifan Carl Guo, Vincent Huang, Jacob Steinhardt, Jacob Andreas
Abstract: Can language models (LMs) learn to faithfully describe their internal computations? Are they better able to describe themselves than other models? We study the extent to which LMs' privileged access to their own internals can be leveraged to produce new techniques for explaining their behavior. Using existing interpretability techniques as a source of ground truth, we fine-tune LMs to generate natural language descriptions of (1) the information encoded by LM features, (2) the causal structure of LMs' internal activations, and (3) the influence of specific input tokens on LM outputs. When trained with only tens of thousands of example explanations, explainer models exhibit non-trivial generalization to new queries. This generalization appears partly attributable to explainer models' privileged access to their own internals: using a model to explain its own computations generally works better than using a *different* model to explain its computations (even if the explainer model is significantly more capable than the target). Our results suggest not only that LMs can learn to reliably explain their internal computations, but that such explanations offer a scalable complement to existing interpretability methods. Code and data at https://github.com/TransluceAI/introspective-interp
Authors: He Zhang, Wenqian Cui, Haoning Xu, Xiaohui Li, Lei Zhu, Haoli Bai, Shaohua Ma, Irwin King
Abstract: Full-Duplex Speech Language Models (FD-SLMs) enable real-time, overlapping conversational interactions, offering a more dynamic user experience compared to traditional half-duplex models. However, existing benchmarks primarily focus on evaluating single-round interactions, neglecting the complexities of multi-round communication. Evaluating FD-SLMs in multi-round settings poses significant challenges, including blurred turn boundaries in communication and context inconsistency during model inference. Also, existing benchmarks often focus solely on evaluating conversational features, neglecting other critical aspects. To address these gaps, we introduce MTR-DuplexBench, a novel benchmark designed for a comprehensive multi-round evaluation of FD-SLMs. MTR-DuplexBench not only segments continuous full-duplex dialogues into discrete turns for turn-by-turn assessment but also incorporates various evaluation aspects, including conversational features, dialogue quality, instruction following, and safety. Experimental results reveal that current FD-SLMs face difficulties in maintaining consistent performance across multiple rounds and evaluation dimensions, highlighting the necessity and effectiveness of our benchmark. The benchmark and code will be available in the future.
Authors: Yaniv Oren, Joery A. de Vries, Pascal R. van der Vaart, Matthijs T. J. Spaan, Wendelin B\"ohmer
Abstract: Model-based reinforcement learning (RL) methods that leverage search are responsible for many milestone breakthroughs in RL. Sequential Monte Carlo (SMC) recently emerged as an alternative to the Monte Carlo Tree Search (MCTS) algorithm which drove these breakthroughs. SMC is easier to parallelize and more suitable to GPU acceleration. However, it also suffers from large variance and path degeneracy which prevent it from scaling well with increased search depth, i.e., increased sequential compute. To address these problems, we introduce Twice Sequential Monte Carlo Tree Search (TSMCTS). Across discrete and continuous environments TSMCTS outperforms the SMC baseline as well as a popular modern version of MCTS as a policy improvement operator, scales favorably with sequential compute, reduces estimator variance and mitigates the effects of path degeneracy while retaining the properties that make SMC natural to parallelize.
Authors: Veeramani Pugazhenthi, Wei-Hsiang Chu, Junwei Lu, Jadyn N. Miyahira, Mahdi Eslamimehr, Pratik Satam, Rozhin Yasaei, Soheil Salehi
Abstract: The use of tiny devices capable of low-latency gesture recognition is gaining momentum in everyday human-computer interaction and especially in medical monitoring fields. Embedded solutions such as fall detection, rehabilitation tracking, and patient supervision require fast and efficient tracking of movements while avoiding unwanted false alarms. This study presents an efficient solution on how to build very efficient motion-based models only using triaxial accelerometer sensors. We explore the capability of the AutoML pipelines to extract the most important features from the data segments. This approach also involves training multiple lightweight machine learning algorithms using the extracted features. We use WeBe Band, a multi-sensor wearable device that is equipped with a powerful enough MCU to effectively perform gesture recognition entirely on the device. Of the models explored, we found that the neural network provided the best balance between accuracy, latency, and memory use. Our results also demonstrate that reliable real-time gesture recognition can be achieved in WeBe Band, with great potential for real-time medical monitoring solutions that require a secure and fast response time.
Authors: Shravan Chaudhari, Yoav Wald, Suchi Saria
Abstract: As we deploy machine learning systems in the real world, a core challenge is to maintain a model that is performant even as the data shifts. Such shifts can take many forms: new classes may emerge that were absent during training, a problem known as open-set recognition, and the distribution of known categories may change. Guarantees on open-set recognition are mostly derived under the assumption that the distribution of known classes, which we call the background distribution, is fixed. In this paper we develop CoLOR, a method that is guaranteed to solve open-set recognition even in the challenging case where the background distribution shifts. We prove that the method works under benign assumptions that the novel class is separable from the non-novel classes, and provide theoretical guarantees that it outperforms a representative baseline in a simplified overparameterized setting. We develop techniques to make CoLOR scalable and robust, and perform comprehensive empirical evaluations on image and text data. The results show that CoLOR significantly outperforms existing open-set recognition methods under background shift. Moreover, we provide new insights into how factors such as the size of the novel class influences performance, an aspect that has not been extensively explored in prior work.
Authors: Haoyang He, Jay Patrikar, Dong-Ki Kim, Max Smith, Daniel McGann, Ali-akbar Agha-mohammadi, Shayegan Omidshafiei, Sebastian Scherer
Abstract: Recent advances in video world modeling have enabled large-scale generative models to simulate embodied environments with high visual fidelity, providing strong priors for prediction, planning, and control. Yet, despite their realism, these models often lack geometric grounding, limiting their use in navigation tasks that require spatial coherence and stability. We introduce Reinforcement Learning with World Grounding (RLWG), a self-supervised post-training framework that aligns pretrained world models with a physically verifiable structure through geometric and perceptual rewards. Analogous to reinforcement learning from verifiable feedback (RLVR) in language models, RLWG can use multiple rewards that measure pose cycle-consistency, depth reprojection, and temporal coherence. We instantiate this framework with GrndCtrl, a reward-aligned adaptation method based on Group Relative Policy Optimization (GRPO), yielding world models that maintain stable trajectories, consistent geometry, and reliable rollouts for embodied navigation. Like post-training alignment in large language models, GrndCtrl leverages verifiable rewards to bridge generative pretraining and grounded behavior, achieving superior spatial coherence and navigation stability over supervised fine-tuning in outdoor environments.
Authors: Wenshuo Wang, Fan Zhang
Abstract: Zero-Shot Super-Resolution Spatiotemporal Forecasting requires a deep learning model to be trained on low-resolution data and deployed for inference on high-resolution. Existing studies consider maintaining similar error across different resolutions as indicative of successful multi-resolution generalization. However, deep learning models serving as alternatives to numerical solvers should reduce error as resolution increases. The fundamental limitation is, the upper bound of physical law frequencies that low-resolution data can represent is constrained by its Nyquist frequency, making it difficult for models to process signals containing unseen frequency components during high-resolution inference. This results in errors being anchored at low resolution, incorrectly interpreted as successful generalization. We define this fundamental phenomenon as a new problem distinct from existing issues: Scale Anchoring. Therefore, we propose architecture-agnostic Frequency Representation Learning. It alleviates Scale Anchoring through resolution-aligned frequency representations and spectral consistency training: on grids with higher Nyquist frequencies, the frequency response in high-frequency bands of FRL-enhanced variants is more stable. This allows errors to decrease with resolution and significantly outperform baselines within our task and resolution range, while incurring only modest computational overhead.
Authors: Yujie Xiao, Qinghao Zhao, Gongzheng Tang, Hao Zhang, Zhuoran Kan, Deyun Zhang, Jun Li, Guangkun Nie, Xiaocheng Fang, Haoyu Wang, Shun Huang, Tong Liu, Jian Liu, Kangyin Chen, Shenda Hong
Abstract: Coronary artery disease (CAD) remains a major global public health burden, yet scalable tools for risk screening are limited. Although coronary computed tomography angiography (CCTA) is a first-line non-invasive diagnostic modality, its widespread use is constrained by resource requirements and radiation exposure. Artificial intelligence--enabled electrocardiography (AI-ECG) may provide a complementary approach for CAD risk stratification. We developed and validated an AI-ECG model using CCTA as the reference standard to estimate severe ($\geq 70\%$) or complete ($\geq 99\%$) stenosis in the four major coronary arteries. In internal validation, the model achieved area under the receiver operating characteristic curve (AUC) values of 0.706--0.744 across vessels and demonstrated consistent performance in external validation (AUCs: 0.673--0.714). Discrimination remained stable among individuals with clinically normal ECGs and across demographic and clinical subgroups. In a dedicated clinical cohort with longitudinal follow-up, vessel-specific risk stratification based on model-predicted probabilities yielded distinct separation between high-risk and low-risk groups in time-to-event analyses using Kaplan--Meier curves, while decision curve analysis suggested potential clinical utility as an adjunctive screening tool. Explainable analyses highlighted waveform patterns associated with elevated predicted risk. These findings support the feasibility of AI-ECG for complementary CAD risk screening and warrant prospective evaluation.
Authors: Honglu Sun, Hao Jing, Zhixiang Dai, Sa Xiao, Wei Xue, Jian Sun, Qifeng Lu
Abstract: A fundamental challenge in numerical weather prediction is to efficiently produce high-resolution forecasts. A common solution is applying downscaling methods, which include dynamical downscaling and statistical downscaling, to the outputs of global models. This work focuses on statistical downscaling, which establishes statistical relationships between low-resolution and high-resolution historical data using statistical models. Deep learning has emerged as a powerful tool for this task, giving rise to various high-performance super-resolution models, which can be directly applied for downscaling, such as diffusion models and Generative Adversarial Networks. This work relies on a diffusion-based downscaling framework named CorrDiff. In contrast to the original work of CorrDiff, the region considered in this work is nearly 40 times larger, and we not only consider surface variables as in the original work, but also encounter high-level variables (six pressure levels) as target downscaling variables. In addition, a global residual connection is added to improve accuracy. In order to generate the 3km forecasts for the China region, we apply our trained models to the 25km global grid forecasts of CMA-GFS, an operational global model of the China Meteorological Administration (CMA), and SFF, a data-driven deep learning-based weather model developed from Spherical Fourier Neural Operators (SFNO). CMA-MESO, a high-resolution regional model, is chosen as the baseline model. The experimental results demonstrate that the forecasts downscaled by our method generally outperform the direct forecasts of CMA-MESO in terms of MAE for the target variables. Our forecasts of radar composite reflectivity show that CorrDiff, as a generative model, can generate fine-scale details that lead to more realistic predictions compared to the corresponding deterministic regression models.
Authors: Guanquan Shi, Haohua Du, Zhiqiang Wang, Xiaoyu Liang, Weiwenpei Liu, Song Bian, Zhenyu Guan
Abstract: Large Language Models (LLMs) are evolving into autonomous agents capable of executing complex workflows via standardized protocols (e.g., MCP). However, this paradigm shifts control from deterministic code to probabilistic inference, creating a fundamental Trust-Authorization Mismatch: static permissions are structurally decoupled from the agent's fluctuating runtime trustworthiness. In this Systematization of Knowledge (SoK), we survey more than 200 representative papers to categorize the emerging landscape of agent security. We propose the Belief-Intention-Permission (B-I-P) framework as a unifying formal lens. By decomposing agent execution into three distinct stages-Belief Formation, Intent Generation, and Permission Grant-we demonstrate that diverse threats, from prompt injection to tool poisoning, share a common root cause: the desynchronization between dynamic trust states and static authorization boundaries. Using the B-I-P lens, we systematically map existing attacks and defenses and identify critical gaps where current mechanisms fail to bridge this mismatch. Finally, we outline a research agenda for shifting from static Role-Based Access Control (RBAC) to dynamic, risk-adaptive authorization.
Authors: Yifan Wu, Jiyue Jiang, Xichen Ye, Yiqi Wang, Chang Zhou, Yitao Xu, Jiayang Chen, He Hu, Weizhong Zhang, Cheng Jin, Jiao Yuan, Yu Li
Abstract: Biological foundation models (BioFMs), pretrained on large-scale biological sequences, have recently shown strong potential in providing meaningful representations for diverse downstream bioinformatics tasks. However, such models often rely on millions to billions of training sequences and billions of parameters, resulting in prohibitive computational costs and significant barriers to reproducibility and accessibility, particularly for academic labs. To address these challenges, we investigate the feasibility of data pruning for BioFM pretraining and propose a post-hoc influence-guided data pruning framework tailored to biological domains. Our approach introduces a subset-based self-influence formulation that enables efficient estimation of sample importance at low computational cost, and builds upon it two simple yet effective selection strategies, namely Top-k Influence (Top I) and Coverage-Centric Influence (CCI). We empirically validate our method on two representative BioFMs, RNA-FM and ESM-C. For RNA, our framework consistently outperforms random selection baselines under an extreme pruning rate of over 99 percent, demonstrating its effectiveness. Furthermore, we show the generalizability of our framework on protein-related tasks using ESM-C. In particular, our coreset even outperforms random subsets that are ten times larger in both RNA and protein settings, revealing substantial redundancy in biological sequence datasets. These findings underscore the potential of influence-guided data pruning to substantially reduce the computational cost of BioFM pretraining, paving the way for more efficient, accessible, and sustainable biological AI research.
Authors: Junchi Lu, Xinke Li, Yuheng Liu, Qi Alfred Chen
Abstract: The increasing use of generative models such as diffusion models for synthetic data augmentation has greatly reduced the cost of data collection and labeling in downstream perception tasks. However, this new data source paradigm may introduce important security concerns. Publicly available generative models are often reused without verification, raising a fundamental question of their safety and trustworthiness. This work investigates backdoor propagation in such emerging generative data supply chain, namely, Data-Chain Backdoor (DCB). Specifically, we find that open-source diffusion models can become hidden carriers of backdoors. Their strong distribution-fitting ability causes them to memorize and reproduce backdoor triggers in generation, which are subsequently inherited by downstream models, resulting in severe security risks. This threat is particularly concerning under clean-label attack scenarios, as it remains effective while having negligible impact on the utility of the synthetic data. We study two attacker choices to obtain a backdoor-carried generator, training from scratch and fine-tuning. While naive fine-tuning leads to weak inheritance of the backdoor, we find that novel designs in the loss objectives and trigger processing can substantially improve the generator's ability to preserve trigger patterns, making fine-tuning a low-cost attack path. We evaluate the effectiveness of DCB under the standard augmentation protocol and further assess data-scarce settings. Across multiple trigger types, we observe that the trigger pattern can be consistently retained in the synthetic data with attack efficacy comparable to the conventional backdoor attack.
Authors: James K Ruffle, Samia Mohinta, Guilherme Pombo, Asthik Biswas, Alan Campbell, Indran Davagnanam, David Doig, Ahmed Hammam, Harpreet Hyare, Farrah Jabeen, Emma Lim, Dermot Mallon, Stephanie Owen, Sophie Wilkinson, Sebastian Brandner, Parashkev Nachev
Abstract: The benefits of artificial intelligence (AI) human partnerships-evaluating how AI agents enhance expert human performance-are increasingly studied. Though rarely evaluated in healthcare, an inverse approach is possible: AI benefiting from the support of an expert human agent. Here, we investigate both human-AI clinical partnership paradigms in the magnetic resonance imaging-guided characterisation of patients with brain tumours. We reveal that human-AI partnerships improve accuracy and metacognitive ability not only for radiologists supported by AI, but also for AI agents supported by radiologists. Moreover, the greatest patient benefit was evident with an AI agent supported by a human one. Synergistic improvements in agent accuracy, metacognitive performance, and inter-rater agreement suggest that AI can create more capable, confident, and consistent clinical agents, whether human or model-based. Our work suggests that the maximal value of AI in healthcare could emerge not from replacing human intelligence, but from AI agents that routinely leverage and amplify it.
Authors: Mozes Jacobs, Thomas Fel, Richard Hakim, Alessandra Brondetta, Demba Ba, T. Andy Keller
Abstract: As Vision Transformers (ViTs) become standard vision backbones, a mechanistic account of their computational phenomenology is essential. Despite architectural cues that hint at dynamical structure, there is no settled framework that interprets Transformer depth as a well-characterized flow. In this work, we introduce the Block-Recurrent Hypothesis (BRH), arguing that trained ViTs admit a block-recurrent depth structure such that the computation of the original $L$ blocks can be accurately rewritten using only $k \ll L$ distinct blocks applied recurrently. Across diverse ViTs, between-layer representational similarity matrices suggest few contiguous phases. To determine whether these phases reflect genuinely reusable computation, we train block-recurrent surrogates of pretrained ViTs: Recurrent Approximations to Phase-structured TransfORmers (Raptor). In small-scale, we demonstrate that stochastic depth and training promote recurrent structure and subsequently correlate with our ability to accurately fit Raptor. We then provide an empirical existence proof for BRH by training a Raptor model to recover $96\%$ of DINOv2 ImageNet-1k linear probe accuracy in only 2 blocks at equivalent computational cost. Finally, we leverage our hypothesis to develop a program of Dynamical Interpretability. We find i) directional convergence into class-dependent angular basins with self-correcting trajectories under small perturbations, ii) token-specific dynamics, where cls executes sharp late reorientations while patch tokens exhibit strong late-stage coherence toward their mean direction, and iii) a collapse to low rank updates in late depth, consistent with convergence to low-dimensional attractors. Altogether, we find a compact recurrent program emerges along ViT depth, pointing to a low-complexity normative solution that enables these models to be studied through principled dynamical systems analysis.
Authors: Marcel Meyer, Sascha Kaltenpoth, Henrik Albers, Kevin Zalipski, Oliver M\"uller
Abstract: Time Series Foundation Models (TSFMs) are transforming the field of forecasting. However, evaluating them on historical data is increasingly difficult due to the risks of train-test sample overlaps and temporal overlaps between correlated train and test time series. To address this, we introduce TS-Arena, a live forecasting platform that shifts evaluation from the known past to the unknown future. Building on the concept of continuous benchmarking, TS-Arena evaluates models on future data. Crucially, we introduce a strict forecasting pre-registration protocol: models must submit predictions before the ground-truth data physically exists. This makes test-set contamination impossible by design. The platform relies on a modular microservice architecture that harmonizes and structures data from different sources and orchestrates containerized model submissions. By enforcing a strict pre-registration protocol on live data streams, TS-Arena prevents information leakage offers a faster alternative to traditional static, infrequently repeated competitions (e.g. the M-Competitions). First empirical results derived from operating TS-Arena over one year of energy time series demonstrate that established TSFMs accumulate robust longitudinal scores over time, while the continuous nature of the benchmark simultaneously allows newcomers to demonstrate immediate competitiveness. TS-Arena provides the necessary infrastructure to assess the true generalization capabilities of modern forecasting models. The platform and corresponding code are available at https://ts-arena.live/.
URLs: https://ts-arena.live/.
Authors: Khalfalla Awedat, Mohamed Abidalrekab, Gurcan Comert, Mustafa Ayad
Abstract: LiDAR-based perception in autonomous systems is constrained by fixed vertical beam resolution and further compromised by beam dropout resulting from environmental occlusions. This paper introduces SuperiorGAT, a graph attention-based framework designed to reconstruct missing elevation information in sparse LiDAR point clouds. By modeling LiDAR scans as beam-aware graphs and incorporating gated residual fusion with feed-forward refinement, SuperiorGAT enables accurate reconstruction without increasing network depth. To evaluate performance, structured beam dropout is simulated by removing every fourth vertical scanning beam. Extensive experiments across diverse KITTI environments, including Person, Road, Campus, and City sequences, demonstrate that SuperiorGAT consistently achieves lower reconstruction error and improved geometric consistency compared to PointNet-based models and deeper GAT baselines. Qualitative X-Z projections further confirm the model's ability to preserve structural integrity with minimal vertical distortion. These results suggest that architectural refinement offers a computationally efficient method for improving LiDAR resolution without requiring additional sensor hardware.
Authors: Jihang Wang, Dongcheng Zhao, Ruolin Chen, Qian Zhang, Yi Zeng
Abstract: Spiking Neural Networks (SNNs) utilize spike-based activations to mimic the brain's energy-efficient information processing. However, the binary and discontinuous nature of spike activations causes vanishing gradients, making adversarial robustness evaluation via gradient descent unreliable. While improved surrogate gradient methods have been proposed, their effectiveness under strong adversarial attacks remains unclear. We propose a more reliable framework for evaluating SNN adversarial robustness. We theoretically analyze the degree of gradient vanishing in surrogate gradients and introduce the Adaptive Sharpness Surrogate Gradient (ASSG), which adaptively evolves the shape of the surrogate function according to the input distribution during attack iterations, thereby enhancing gradient accuracy while mitigating gradient vanishing. In addition, we design an adversarial attack with adaptive step size under the $L_\infty$ constraint-Stable Adaptive Projected Gradient Descent (SA-PGD), achieving faster and more stable convergence under imprecise gradients. Extensive experiments show that our approach substantially increases attack success rates across diverse adversarial training schemes, SNN architectures and neuron models, providing a more generalized and reliable evaluation of SNN adversarial robustness. The experimental results further reveal that the robustness of current SNNs has been significantly overestimated and highlighting the need for more dependable adversarial training methods. The code is released at https://github.com/craree/ASSG-SNNs-Robustness-Evaluation
URLs: https://github.com/craree/ASSG-SNNs-Robustness-Evaluation
Authors: Yingru Li, Jiacai Liu, Jiawei Xu, Yuxuan Tong, Ziniu Li, Qian Liu, Baoxiang Wang
Abstract: Policy gradient methods for Large Language Models optimize a policy $\pi_\theta$ via a surrogate objective computed from samples of a rollout policy $\pi_{\text{roll}}$. However, modern LLM-RL pipelines suffer from unavoidable implementation divergences -- backend discrepancies, Mixture-of-Experts routing discontinuities, and distributed training staleness -- causing off-policy mismatch ($\pi_{\text{roll}} \neq \pi_\theta$) and approximation errors between the surrogate and the true objective. We demonstrate that classical trust region bounds on this error scale as $O(T^2)$ with sequence length $T$, rendering them vacuous for long-horizon tasks. To address this, we derive a family of bounds -- both KL-based and TV-based -- including a Pinsker-Marginal bound ($O(T^{3/2})$), a Mixed bound ($O(T)$), and an Adaptive bound that strictly generalizes the Pinsker-Marginal bound via per-position importance-ratio decomposition. Taking the minimum over all bounds yields the tightest known guarantee across all divergence regimes. Crucially, all bounds depend on the maximum token-level divergence $D_{\mathrm{KL}}^{\mathrm{tok,max}}$ (or $D_{\mathrm{TV}}^{\mathrm{tok,max}}$), a sequence-level quantity that cannot be controlled by token-independent methods like PPO clipping. We propose Trust Region Masking (TRM), which masks entire sequences violating the trust region, enabling the first non-vacuous monotonic improvement guarantees for long-horizon LLM-RL.
Authors: Cheonkam Jeong, Adeline Nyamathi
Abstract: Despite strong recent progress in Emotion Recognition in Conversation (ERC), two gaps remain: we lack clear understanding of which modeling choices materially affect performance, and we have limited linguistic analysis linking recognition findings to actionable generation cues. We address both via a systematic study on IEMOCAP. For recognition, we conduct controlled ablations with 10 random seeds and paired tests (with correction for multiple comparisons), yielding three findings. First, conversational context is dominant: performance saturates quickly, with roughly 90% of gain achieved using only the most recent 10-30 preceding turns. Second, hierarchical sentence representations improve utterance-only recognition (K=0), but the benefit vanishes once turn-level context is available, suggesting conversational history subsumes intra-utterance structure. Third, external affective lexicon (SenticNet) integration does not improve results, consistent with pretrained encoders already capturing affective signal. Under strictly causal (past-only) setting, our simple models attain strong performance (82.69% 4-way; 67.07% 6-way weighted F1). For linguistic analysis, we examine 5,286 discourse-marker occurrences and find reliable association between emotion and marker position (p < 0.0001). Sad utterances show reduced left-periphery marker usage (21.9%) relative to other emotions (28-32%), aligning with accounts linking left-periphery markers to active discourse management. This pattern is consistent with Sad benefiting most from conversational context (+22%p), suggesting sadness relies more on discourse history than overt pragmatic signaling.
Authors: Elchanan Mossel
Abstract: Recent reports claim that Large Language Models (LLMs) have achieved the ability to derive new science and exhibit human-level general intelligence. We argue that such claims are not rigorous scientific claims, as they do not satisfy Popper's refutability principle (often termed falsifiability), which requires that scientific statements be capable of being disproven. We identify several methodological pitfalls in current AI research on reasoning, including the inability to verify the novelty of findings due to opaque and non-searchable training data, the lack of reproducibility caused by continuous model updates, and the omission of human-interaction transcripts, which obscures the true source of scientific discovery. Additionally, the absence of counterfactuals and data on failed attempts creates a selection bias that may exaggerate LLM capabilities. To address these challenges, we propose guidelines for scientific transparency and reproducibility for research on reasoning by LLMs. Establishing such guidelines is crucial for both scientific integrity and the ongoing societal debates regarding fair data usage.
Authors: Shanglin Li, Shiwen Chu, Okan Ko\c{c}, Yi Ding, Qibin Zhao, Motoaki Kawanabe, Ziheng Chen
Abstract: Electroencephalography (EEG)-based brain-computer interfaces facilitate direct communication with a computer, enabling promising applications in human-computer interactions. However, their utility is currently limited because EEG decoding often suffers from poor generalization due to distribution shifts across domains (e.g., subjects). Learning robust representations that capture underlying task-relevant information would mitigate these shifts and improve generalization. One promising approach is to exploit the underlying hierarchical structure in EEG, as recent studies suggest that hierarchical cognitive processes, such as visual processing, can be encoded in EEG. While many decoding methods still rely on Euclidean embeddings, recent work has begun exploring hyperbolic geometry for EEG. Hyperbolic spaces, regarded as the continuous analogue of tree structures, provide a natural geometry for representing hierarchical data. In this study, we first empirically demonstrate that EEG data exhibit hyperbolicity and show that hyperbolic embeddings improve generalization. Motivated by these findings, we propose HEEGNet, a hybrid hyperbolic network architecture to capture the hierarchical structure in EEG and learn domain-invariant hyperbolic embeddings. To this end, HEEGNet combines both Euclidean and hyperbolic encoders and employs a novel coarse-to-fine domain adaptation strategy. Extensive experiments on multiple public EEG datasets, covering visual evoked potentials, emotion recognition, and intracranial EEG, demonstrate that HEEGNet achieves state-of-the-art performance. The code is available at https://github.com/fightlesliefigt/HEEGNet
Authors: Christophe Bonneville, Nathan Bieberdorf, Pieterjan Robbe, Mark Asta, Habib Najm, Laurent Capolungo, Cosmin Safta
Abstract: Phase-field simulations of liquid metal dealloying (LMD) can capture complex microstructural evolutions but can be prohibitively expensive for large domains and long time horizons. In this paper, we introduce a fully convolutional, conditionally parameterized U-Net surrogate designed to extrapolate far beyond its training data in both space and time. The architecture integrates convolutional self-attention, physically informed padding, and a flood-fill corrector method to maintain accuracy under extreme extrapolation, while conditioning on simulation parameters allows for flexible time-step skipping and adaptation to varying alloy compositions. To remove the need for costly solver-based initialization, we couple the surrogate with a conditional diffusion model that generates synthetic, physically consistent initial conditions. We train our surrogate on simulations generated over small domain sizes and short time spans, but, by taking advantage of the convolutional nature of U-Nets, we are able to run and extrapolate surrogate simulations for longer time horizons than what would be achievable with classic numerical solvers. Across multiple alloy compositions, the framework is able to reproduce the LMD physics accurately. It predicts key quantities of interest and spatial statistics with relative errors typically below 5% in the training regime and under 15% during large-scale, long time-horizon extrapolations. Our framework can also deliver speed-ups of up to 36,000 times, bringing the time to run weeks-long simulations down to a few seconds. This work is a first stepping stone towards high-fidelity extrapolation in both space and time of phase-field simulation for LMD.
Authors: Xiaoyu Ma, David Patterson
Abstract: Large Language Model (LLM) inference is hard. The autoregressive Decode phase of the underlying Transformer model makes LLM inference fundamentally different from training. Exacerbated by recent AI trends, the primary challenges are memory and interconnect rather than compute. To address these challenges, we highlight four architecture research opportunities: High Bandwidth Flash for 10X memory capacity with HBM-like bandwidth; Processing-Near-Memory and 3D memory-logic stacking for high memory bandwidth; and low-latency interconnect to speedup communication. While our focus is datacenter AI, we also review their applicability for mobile devices.
Authors: Wonhyeok Choi, Shutong Ding, Minwoo Choi, Jungwan Woo, Kyumin Hwang, Jaeyeul Kim, Ye Shi, Sunghoon Im
Abstract: Diffusion policies have emerged as a powerful approach for robotic control, demonstrating superior expressiveness in modeling multimodal action distributions compared to conventional policy networks. However, their integration with online reinforcement learning remains challenging due to fundamental incompatibilities between diffusion model training objectives and standard RL policy improvement mechanisms. This paper presents the first comprehensive review and empirical analysis of current Online Diffusion Policy Reinforcement Learning (Online DPRL) algorithms for scalable robotic control systems. We propose a novel taxonomy that categorizes existing approaches into four distinct families--Action-Gradient, Q-Weighting, Proximity-Based, and Backpropagation Through Time (BPTT) methods--based on their policy improvement mechanisms. Through extensive experiments on a unified NVIDIA Isaac Lab benchmark encompassing 12 diverse robotic tasks, we systematically evaluate representative algorithms across five critical dimensions: task diversity, parallelization capability, diffusion step scalability, cross-embodiment generalization, and environmental robustness. Our analysis identifies key findings regarding the fundamental trade-offs inherent in each algorithmic family, particularly concerning sample efficiency and scalability. Furthermore, we reveal critical computational and algorithmic bottlenecks that currently limit the practical deployment of online DPRL. Based on these findings, we provide concrete guidelines for algorithm selection tailored to specific operational constraints and outline promising future research directions to advance the field toward more general and scalable robotic learning systems.
Authors: Peiyuan Jing, Yue Yang, Chun-Wun Cheng, Zhenxuan Zhang, Liutao Yang, Thiago V. Lima, Klaus Strobel, Antoine Leimgruber, Angelica Aviles-Rivero, Guang Yang, Javier Montoya
Abstract: Low-dose Positron Emission Tomography (PET) imaging reduces patient radiation exposure but suffers from increased noise that degrades image quality and diagnostic reliability. Although diffusion models have demonstrated strong denoising capability, their stochastic nature makes it challenging to enforce anatomically consistent structures, particularly in low signal-to-noise regimes and volumetric whole-body imaging. We propose Wavelet-Conditioned ControlNet (WCC-Net), a fully 3D diffusion-based framework that introduces explicit frequency-domain structural priors via wavelet representations to guide volumetric PET denoising. By injecting wavelet-based structural guidance into a frozen pretrained diffusion backbone through a lightweight control branch, WCC-Net decouples anatomical structure from noise while preserving generative expressiveness and 3D structural continuity. Extensive experiments demonstrate that WCC-Net consistently outperforms CNN-, GAN-, and diffusion-based baselines. On the internal 1/20-dose test set, WCC-Net improves PSNR by +1.21 dB and SSIM by +0.008 over a strong diffusion baseline, while reducing structural distortion (GMSD) and intensity error (NMAE). Moreover, WCC-Net generalizes robustly to unseen dose levels (1/50 and 1/4), achieving superior quantitative performance and improved volumetric anatomical consistency.
Authors: Yan Wang, Sayeef Abdullah, Partho Hassan, Sabit Hassan
Abstract: The dataset spans diverse artistic styles, including regionally grounded aesthetics from the Middle East, Northern Europe, East Asia, and South Asia, alongside general categories such as sketch and oil painting. All images are generated using the Moonworks Lunara model and intentionally crafted to embody distinct, high-quality aesthetic styles, yielding a first-of-its-kind dataset with substantially higher aesthetic scores, exceeding even aesthetics-focused datasets, and general-purpose datasets by a larger margin. Each image is accompanied by a human-refined prompt and structured annotations that jointly describe salient objects, attributes, relationships, and stylistic cues. Unlike large-scale web-derived datasets that emphasize breadth over precision, the Lunara Aesthetic Dataset prioritizes aesthetic quality, stylistic diversity, and licensing transparency, and is released under the Apache 2.0 license to support research and unrestricted academic and commercial use.
Authors: Gyu-Il Kim, Dae-Won Kim, Jaesung Lee
Abstract: Unsupervised feature selection aims to identify a compact subset of features that captures the intrinsic structure of data without supervised label. Most existing studies evaluate the performance of methods using the single-label dataset that can be instantiated by selecting a label from multi-label data while maintaining the original features. Because the chosen label can vary arbitrarily depending on the experimental setting, the superiority among compared methods can be changed with regard to which label happens to be selected. Thus, evaluating unsupervised feature selection methods based solely on single-label accuracy is unreasonable for assessing their true discriminative ability. This study revisits this evaluation paradigm by adopting a multi-label classification framework. Experiments on 21 multi-label datasets using several representative methods demonstrate that performance rankings differ markedly from those reported under single-label settings, suggesting the possibility of multi-label evaluation settings for fair and reliable comparison of unsupervised feature selection methods.
Authors: Ritik Raina, Abe Leite, Alexandros Graikos, Seoyoung Ahn, Dimitris Samaras, Gregory J. Zelinsky
Abstract: Human vision combines low-resolution "gist" information from the visual periphery with sparse but high-resolution information from fixated locations to construct a coherent understanding of a visual scene. In this paper, we introduce MetamerGen, a tool for generating scenes that are aligned with latent human scene representations. MetamerGen is a latent diffusion model that combines peripherally obtained scene gist information with information obtained from scene-viewing fixations to generate image metamers for what humans understand after viewing a scene. Generating images from both high and low resolution (i.e. "foveated") inputs constitutes a novel image-to-image synthesis problem, which we tackle by introducing a dual-stream representation of the foveated scenes consisting of DINOv2 tokens that fuse detailed features from fixated areas with peripherally degraded features capturing scene context. To evaluate the perceptual alignment of MetamerGen generated images to latent human scene representations, we conducted a same-different behavioral experiment where participants were asked for a "same" or "different" response between the generated and the original image. With that, we identify scene generations that are indeed metamers for the latent scene representations formed by the viewers. MetamerGen is a powerful tool for understanding scene understanding. Our proof-of-concept analyses uncovered specific features at multiple levels of visual processing that contributed to human judgments. While it can generate metamers even conditioned on random fixations, we find that high-level semantic alignment most strongly predicts metamerism when the generated scenes are conditioned on viewers' own fixated regions.
Authors: Deming Chen, Vijay Ganesh, Weikai Li, Yingyan Celine Lin, Yong Liu, Subhasish Mitra, David Z. Pan, Ruchir Puri, Jason Cong, Yizhou Sun
Abstract: This report distills the discussions and recommendations from the NSF Workshop on AI for Electronic Design Automation (EDA), held on December 10, 2024 in Vancouver alongside NeurIPS 2024. Bringing together experts across machine learning and EDA, the workshop examined how AI-spanning large language models (LLMs), graph neural networks (GNNs), reinforcement learning (RL), neurosymbolic methods, etc.-can facilitate EDA and shorten design turnaround. The workshop includes four themes: (1) AI for physical synthesis and design for manufacturing (DFM), discussing challenges in physical manufacturing process and potential AI applications; (2) AI for high-level and logic-level synthesis (HLS/LLS), covering pragma insertion, program transformation, RTL code generation, etc.; (3) AI toolbox for optimization and design, discussing frontier AI developments that could potentially be applied to EDA tasks; and (4) AI for test and verification, including LLM-assisted verification tools, ML-augmented SAT solving, security/reliability challenges, etc. The report recommends NSF to foster AI/EDA collaboration, invest in foundational AI for EDA, develop robust data infrastructures, promote scalable compute infrastructure, and invest in workforce development to democratize hardware design and enable next-generation hardware systems. The workshop information can be found on the website https://ai4eda-workshop.github.io/.
Authors: Hendrika Maclean, Mert Can Cakmak, Muzakkiruddin Ahmed Mohammed, Shames Al Mandalawi, John Talburt
Abstract: Large language models are now used daily for writing, search, and analysis, and their natural language understanding continues to improve. However, they remain unreliable on exact numerical calculation and on producing outputs that are straightforward to audit. We study synthetic payroll system as a focused, high-stakes example and evaluate whether models can understand a payroll schema, apply rules in the right order, and deliver cent-accurate results. Our experiments span a tiered dataset from basic to complex cases, a spectrum of prompts from minimal baselines to schema-guided and reasoning variants, and multiple model families including GPT, Claude, Perplexity, Grok and Gemini. Results indicate clear regimes where careful prompting is sufficient and regimes where explicit computation is required. The work offers a compact, reproducible framework and practical guidance for deploying LLMs in settings that demand both accuracy and assurance.
Authors: Trong Khiem Tran, Manh Cuong Dao, Phi Le Nguyen, Thao Nguyen Truong, Trong Nghia Hoang
Abstract: Adapting pre-trained models to unseen feature modalities has become increasingly important due to the growing need for cross-disciplinary knowledge integration. A key challenge here is how to align the representation of new modalities with the most relevant parts of the pre-trained model's representation space to enable accurate knowledge transfer. This requires combining feature alignment with target fine-tuning, but uncalibrated combinations can exacerbate misalignment between the source and target feature-label structures and reduce target generalization. Existing work, however, lacks a theoretical understanding of this critical interaction between feature alignment and target fitting. To bridge this gap, we develop a principled framework that establishes a provable generalization bound on the target error, which explains the interaction between feature alignment and target fitting through a novel concept of feature-label distortion. This bound offers actionable insights into how this interaction should be optimized for practical algorithm design. The resulting approach achieves significantly improved performance over state-of-the-art methods across a wide range of benchmark datasets.
Authors: Yingxiao Huo, Satya Prakash Dash, Radu Stoican, Samuel Kaski, Mingfei Sun
Abstract: Natural gradients have long been studied in deep reinforcement learning due to their fast convergence properties and covariant weight updates. However, computing natural gradients requires inversion of the Fisher Information Matrix (FIM) at each iteration, which is computationally prohibitive in nature. In this paper, we present an efficient and scalable natural policy optimization technique that leverages a rank-1 approximation to full inverse-FIM. We theoretically show that under certain conditions, a rank-1 approximation to inverse-FIM converges faster than policy gradients and, under some conditions, enjoys the same sample complexity as stochastic policy gradient methods. We benchmark our method on a diverse set of environments and show that it achieves superior performance to standard actor-critic and trust-region baselines.
Authors: Hansheng Ren
Abstract: The pursuit of scale in deep learning has entrenched a trade-off: computational throughput is prioritized at the expense of numerical precision. We argue this compromise is fundamentally at odds with the requirements of general intelligence. We propose the \textit{Exactness Hypothesis}: high-order causal reasoning -- a cornerstone of AGI -- demands a substrate supporting arbitrary-precision, logically consistent arithmetic. We trace prevalent LLM failures, such as logical hallucinations and incoherence, to the inherent limitations of IEEE 754 floating-point arithmetic, where approximation errors compound catastrophically in deep functions. As a solution, we present the Halo Architecture, which transitions the computational foundation from approximate reals ($\mathbb{R}$) to exact rationals ($\mathbb{Q}$). Halo is realized through a custom Exact Inference Unit (EIU), whose design -- featuring asynchronous MIMD reduction and dual-modular redundancy -- resolves the performance and reliability bottlenecks of exact computation at scale. Crucially, we theoretically prove that Halo's number-theoretic quantization yields a quadratic error decay ($\mathcal{O}(D^{-2})$), strictly superior to the linear barrier of standard fixed-point arithmetic. In rigorous simulations, 600B-parameter BF16 models fail in chaotic systems within steps, while Halo sustains perfect numerical fidelity indefinitely. Our work posits exact arithmetic as non-negotiable for advancing reasoning-capable AGI and provides a co-designed hardware-software path toward verifiable, exascale-ready AI systems
Authors: Seokhyeon Park, Soohyun Lee, Eugene Choi, Hyunwoo Kim, Minkyu Kweon, Yumin Song, Jinwook Seo
Abstract: While generative AI enables high-fidelity UI generation from text prompts, users struggle to articulate design intent and evaluate or refine results-creating gulfs of execution and evaluation. To understand the information needed for UI generation, we conducted a thematic analysis of UI prompting guidelines, identifying key design semantics and discovering that they are hierarchical and interdependent. Leveraging these findings, we developed a system that enables users to specify semantics, visualize relationships, and extract how semantics are reflected in generated UIs. By making semantics serve as an intermediate representation between human intent and AI output, our system bridges both gulfs by making requirements explicit and outcomes interpretable. A comparative user study suggests that our approach enhances users' perceived control over intent expression and outcome interpretation, and facilitates more predictable iterative refinement. Our work demonstrates how explicit semantic representation enables systematic and explainable exploration of design possibilities in AI-driven UI design.
Authors: Anower Zihad, Felix Owino, Ming Tang, Chao Huang
Abstract: Split learning is a distributed training paradigm where a neural network is partitioned between clients and a server, which allows data to remain at the client while only intermediate activations are shared. Traditional split learning relies on end-to-end backpropagation across the client-server split point. This incurs a large communication overhead (i.e., forward activations and backward gradients need to be exchanged every iteration) and significant memory use (for storing activations and gradients). In this paper, we develop a beyond-backpropagation training method for split learning. In this approach, the client and server train their model partitions semi-independently, using local loss signals instead of propagated gradients. In particular, the client's network is augmented with a small auxiliary classifier at the split point to provide a local error signal, while the server trains on the client's transmitted activations using the true loss function. This decoupling removes the need to send backward gradients, which cuts communication costs roughly in half and also reduces memory overhead (as each side only stores local activations for its own backward pass). We evaluate our approach on CIFAR-10 and CIFAR-100. Our experiments show two key results. First, the proposed approach achieves performance on par with standard split learning that uses backpropagation. Second, it significantly reduces communication (of transmitting activations/gradient) by 50% and peak memory usage by up to 58%.
Authors: Manuel Wendl, Yarden As, Manish Prajapat, Anton Pollak, Stelian Coros, Andreas Krause
Abstract: Safe exploration is a key requirement for reinforcement learning (RL) agents to learn and adapt online, beyond controlled (e.g. simulated) environments. In this work, we tackle this challenge by utilizing suboptimal yet conservative policies (e.g., obtained from offline data or simulators) as priors. Our approach, SOOPER, uses probabilistic dynamics models to optimistically explore, yet pessimistically fall back to the conservative policy prior if needed. We prove that SOOPER guarantees safety throughout learning, and establish convergence to an optimal policy by bounding its cumulative regret. Extensive experiments on key safe RL benchmarks and real-world hardware demonstrate that SOOPER is scalable, outperforms the state-of-the-art and validate our theoretical guarantees in practice.
Authors: Kei Saito
Abstract: Large language models exhibit a systematic tendency toward early semantic commitment: given ambiguous input, they collapse multiple valid interpretations into a single response before sufficient context is available. We present a formal framework for text-to-state mapping ($\phi: \mathcal{T} \to \mathcal{S}$) that transforms natural language into a non-collapsing state space where multiple interpretations coexist. The mapping decomposes into three stages: conflict detection, interpretation extraction, and state construction. We instantiate $\phi$ with a hybrid extraction pipeline combining rule-based segmentation for explicit conflict markers (adversative conjunctions, hedging expressions) with LLM-based enumeration of implicit ambiguity (epistemic, lexical, structural). On a test set of 68 ambiguous sentences, the resulting states preserve interpretive multiplicity: mean state entropy $H = 1.087$ bits across ambiguity categories, compared to $H = 0$ for collapse-based baselines. We additionally instantiate the rule-based conflict detector for Japanese markers to illustrate cross-lingual portability. This framework extends Non-Resolution Reasoning (NRR) by providing the missing algorithmic bridge between text and the NRR state space, enabling architectural collapse deferment in LLM inference. Design principles for state-to-state transformations are detailed in the Appendix, with empirical validation on 580 test cases showing 0% collapse for principle-satisfying operators versus up to 17.8% for violating operators.
Authors: Yuetian Chen, Kaiyuan Zhang, Yuntao Du, Edoardo Stoppa, Charles Fleming, Ashish Kundu, Bruno Ribeiro, Ninghui Li
Abstract: Diffusion Language Models (DLMs) represent a promising alternative to autoregressive language models, using bidirectional masked token prediction. Yet their susceptibility to privacy leakage via Membership Inference Attacks (MIA) remains critically underexplored. This paper presents the first systematic investigation of MIA vulnerabilities in DLMs. Unlike the autoregressive models' single fixed prediction pattern, DLMs' multiple maskable configurations exponentially increase attack opportunities. This ability to probe many independent masks dramatically improves detection chances. To exploit this, we introduce SAMA (Subset-Aggregated Membership Attack), which addresses the sparse signal challenge through robust aggregation. SAMA samples masked subsets across progressive densities and applies sign-based statistics that remain effective despite heavy-tailed noise. Through inverse-weighted aggregation prioritizing sparse masks' cleaner signals, SAMA transforms sparse memorization detection into a robust voting mechanism. Experiments on nine datasets show SAMA achieves 30% relative AUC improvement over the best baseline, with up to 8 times improvement at low false positive rates. These findings reveal significant, previously unknown vulnerabilities in DLMs, necessitating the development of tailored privacy defenses.
Authors: Kaito Baba, Yoshihiko Ozaki, Shuhei Watanabe
Abstract: We propose conditional PED-ANOVA (condPED-ANOVA), a principled framework for estimating hyperparameter importance (HPI) in conditional search spaces, where the presence or domain of a hyperparameter can depend on other hyperparameters. Although the original PED-ANOVA provides a fast and efficient way to estimate HPI within the top-performing regions of the search space, it assumes a fixed, unconditional search space and therefore cannot properly handle conditional hyperparameters. To address this, we introduce a conditional HPI for top-performing regions and derive a closed-form estimator that accurately reflects conditional activation and domain changes. Experiments show that naive adaptations of existing HPI estimators yield misleading or uninterpretable importances in conditional settings, whereas condPED-ANOVA consistently provides meaningful importances that reflect the underlying conditional structure. Our code is publicly available at https://github.com/kAIto47802/condPED-ANOVA.
Authors: Ruicheng Ao, David Simchi-Levi, Xinshang Wang
Abstract: Operations Research practitioners routinely debug infeasible models through an iterative process: analyzing Irreducible Infeasible Subsystems (\IIS{}), identifying constraint conflicts, and systematically repairing formulations until feasibility is achieved. Yet existing LLM benchmarks evaluate OR as one-shot translation -- given a problem description, generate solver code -- ignoring this diagnostic loop entirely. We introduce two benchmarks that place the \textbf{solver in the evaluation loop}. \textbf{\ORDebug{}} evaluates iterative self-correction through 5,000+ problems spanning 9 error types; each repair action triggers solver re-execution and \IIS{} recomputation, providing deterministic, verifiable feedback. \textbf{\ORBias{}} evaluates behavioral rationality through 2,000 newsvendor instances (1,000 ID + 1,000 OOD), measuring systematic deviations from closed-form optimal policies. Across 26 models and 12,000+ samples, we find that domain-specific RLVR training enables an 8B model to surpass frontier APIs: 95.3\% vs 86.2\% recovery rate (+9.1\%), 62.4\% vs 47.8\% diagnostic accuracy (+14.6\%), and 2.25 vs 3.78 steps to resolution (1.7$\times$ faster). On \ORBias{}, curriculum training achieves the only negative ID$\rightarrow$OOD bias drift among models evaluated (-9.6\%), reducing systematic bias by 48\% (from 20.0\% to 10.4\%). These results demonstrate that process-level evaluation with verifiable oracles enables targeted training that outperforms scale.
Authors: Yu Xie, Xing Kai Ren, Ying Qi, Hu Yao
Abstract: While works such as OneRec have validated the scaling laws of Large Language Models (LLMs) in recommender systems, they rely on a cumbersome separate vocabulary. This dependency prevents the model architecture from reusing native LLM vocabularies, resulting in high maintenance costs and poor scalability. In response, we aim to efficiently reuse open-source LLM architectures without constructing a separate tokenization vocabulary. Furthermore, we identify that the optimization strategy of OneRec Gradient Bounded Policy Optimization (GBPO),suffers from a "Symmetric Conservatism" problem: its static gradient boundaries structurally suppress the update momentum required for cold-start items and fail to prevent diversity collapse in high-noise environments.To address this issue, we propose SAGE (Sequence-level Adaptive Gradient Evolution), a unified optimization framework tailored for list-wise generative recommendation. SAGE introduces two key innovations:(1) Sequence-level Signal Decoupling: By combining a geometric mean importance ratio with decoupled multi-objective advantages, we eliminate token-level variance and resolve the "Reward Collapse" problem. (2) Asymmetric Adaptive Dynamics: We construct a dynamic gradient manifold that applies a "Boost Factor" to high-potential cold start items to achieve super-linear updates and employs an "Entropy Aware Penalty" to break information cocoons. Theoretical analysis and empirical results demonstrate that SAGE effectively unblocks cold-start traffic and sustains recommendation diversity, all while retaining the numerical stability of GBPO.
Authors: Chengyu Du, Xintao Wang, Aili Chen, Weiyuan Li, Rui Xu, Junteng Liu, Zishan Huang, Rong Tian, Zijun Sun, Yuhao Li, Liheng Feng, Deming Ding, Pengyu Zhao, Yanghua Xiao
Abstract: LLM role-playing, i.e., using LLMs to simulate specific personas, has emerged as a key capability in various applications, such as companionship, content creation, and digital games. While current models effectively capture character tones and knowledge, simulating the inner thoughts behind their behaviors remains a challenge. Towards cognitive simulation in LLM role-play, previous efforts mainly suffer from two deficiencies: data with high-quality reasoning traces, and reliable reward signals aligned with human preferences. In this paper, we propose HER, a unified framework for cognitive-level persona simulation. HER introduces dual-layer thinking, which distinguishes characters' first-person thinking from LLMs' third-person thinking. To bridge these gaps, we curate reasoning-augmented role-playing data via reverse engineering and construct human-aligned principles and reward models. Leveraging these resources, we train HER models based on Qwen3-32B via supervised and reinforcement learning. Extensive experiments validate the effectiveness of our approach. Notably, our models significantly outperform the Qwen3-32B baseline, achieving a 30.26 improvement on the CoSER benchmark and a 14.97 gain on the Minimax Role-Play Bench. Our datasets, principles, and models will be released to facilitate future research.
Authors: Chao Wang, Licheng Jiao, Lingling Li, Jiaxuan Zhao, Guanchun Wang, Fang Liu, Shuyuan Yang
Abstract: Handcrafted optimizers become prohibitively inefficient for complex black-box optimization (BBO) tasks. MetaBBO addresses this challenge by meta-learning to automatically configure optimizers for low-level BBO tasks, thereby eliminating heuristic dependencies. However, existing methods typically require extensive handcrafted training tasks to learn meta-strategies that generalize to target tasks, which poses a critical limitation for realistic applications with unknown task distributions. To overcome the issue, we propose the Adaptive meta Black-box Optimization Model (ABOM), which performs online parameter adaptation using solely optimization data from the target task, obviating the need for predefined task distributions. Unlike conventional metaBBO frameworks that decouple meta-training and optimization phases, ABOM introduces a closed-loop adaptive parameter learning mechanism, where parameterized evolutionary operators continuously self-update by leveraging generated populations during optimization. This paradigm shift enables zero-shot optimization: ABOM achieves competitive performance on synthetic BBO benchmarks and realistic unmanned aerial vehicle path planning problems without any handcrafted training tasks. Visualization studies reveal that parameterized evolutionary operators exhibit statistically significant search patterns, including natural selection and genetic recombination.
Authors: Catherine Yeh, Anh Truong, Mira Dontcheva, Bryan Wang
Abstract: Video storytelling is often constrained by available material, limiting creative expression and leaving undesired narrative gaps. Generative video offers a new way to address these limitations by augmenting captured media with tailored visuals. To explore this potential, we interviewed eight video creators to identify opportunities and challenges in integrating generative video into their workflows. Building on these insights and established filmmaking principles, we developed Vidmento, a tool for authoring hybrid video stories that combine captured and generated media through context-aware expansion. Vidmento surfaces opportunities for story development, generates clips that blend stylistically and narratively with surrounding media, and provides controls for refinement. In a study with 12 creators, Vidmento supported narrative development and exploration by systematically expanding initial materials with generative media, enabling expressive video storytelling aligned with creative intent. We highlight how creators bridge story gaps with generative content and where they find this blending capability most valuable.
Authors: Tanay Raghunandan Srinivasa, Vivek Deulkar, Jia Bhargava, Mohammad Hajiesmaili, Prashant Shenoy
Abstract: Battery energy storage systems are increasingly deployed as fast-responding resources for grid balancing services such as frequency regulation and for mitigating renewable generation uncertainty. However, repeated charging and discharging induces cycling degradation and reduces battery lifetime. This paper studies the real-time scheduling of a heterogeneous battery fleet that collectively tracks a stochastic balancing signal subject to per-battery ramp-rate and capacity constraints, while minimizing long-term cycling degradation. Cycling degradation is fundamentally path-dependent: it is determined by charge-discharge cycles formed by the state-of-charge (SoC) trajectory and is commonly quantified via rainflow cycle counting. This non-Markovian structure makes it difficult to express degradation as an additive per-time-step cost, complicating classical dynamic programming approaches. We address this challenge by formulating the fleet scheduling problem as a Markov decision process (MDP) with constrained action space and designing a dense proxy reward that provides informative feedback at each time step while remaining aligned with long-term cycle-depth reduction. To scale learning to large state-action spaces induced by fine-grained SoC discretization and asymmetric per-battery constraints, we develop a function-approximation reinforcement learning method using an Extreme Learning Machine (ELM) as a random nonlinear feature map combined with linear temporal-difference learning. We evaluate the proposed approach on a toy Markovian signal model and on a Markovian model trained from real-world regulation signal traces obtained from the University of Delaware, and demonstrate consistent reductions in cycle-depth occurrence and degradation metrics compared to baseline scheduling policies.
Authors: Tanay Raghunandan Srinivasa, Vivek Deulkar, Aviruch Bhatia, Vishal Garg
Abstract: We study the joint operation and sizing of cooling infrastructure for commercial HVAC systems using reinforcement learning, with the objective of minimizing life-cycle cost over a 30-year horizon. The cooling system consists of a fixed-capacity electric chiller and a thermal energy storage (TES) unit, jointly operated to meet stochastic hourly cooling demands under time-varying electricity prices. The life-cycle cost accounts for both capital expenditure and discounted operating cost, including electricity consumption and maintenance. A key challenge arises from the strong asymmetry in capital costs: increasing chiller capacity by one unit is far more expensive than an equivalent increase in TES capacity. As a result, identifying the right combination of chiller and TES sizes, while ensuring zero loss-of-cooling-load under optimal operation, is a non-trivial co-design problem. To address this, we formulate the chiller operation problem for a fixed infrastructure configuration as a finite-horizon Markov Decision Process (MDP), in which the control action is the chiller part-load ratio (PLR). The MDP is solved using a Deep Q Network (DQN) with a constrained action space. The learned DQN RL policy minimizes electricity cost over historical traces of cooling demand and electricity prices. For each candidate chiller-TES sizing configuration, the trained policy is evaluated. We then restrict attention to configurations that fully satisfy the cooling demand and perform a life-cycle cost minimization over this feasible set to identify the cost-optimal infrastructure design. Using this approach, we determine the optimal chiller and thermal energy storage capacities to be 700 and 1500, respectively.
Authors: Sophia N. Wilson, Sebastian Mair, Mophat Okinyi, Erik B. Dam, Janin Koch, Raghavendra Selvan
Abstract: Large-scale data has fuelled the success of frontier artificial intelligence (AI) models over the past decade. This expansion has relied on sustained efforts by large technology corporations to aggregate and curate internet-scale datasets. In this work, we examine the environmental, social, and economic costs of large-scale data in AI through a sustainability lens. We argue that the field is shifting from building models from data to actively creating data for building models. We characterise this transition as hyper-datafication, which marks a critical juncture for the future of frontier AI and its societal impacts. To quantify and contextualise data-related costs, we analyse approximately 550,000 datasets from the Hugging Face Hub, focusing on dataset growth, storage-related energy consumption and carbon footprint, and societal representation using language data. We complement this analysis with qualitative responses from data workers in Kenya to examine the labour involved, including direct employment by big tech corporations and exposure to graphic content. We further draw on external data sources to substantiate our findings by illustrating the global disparity in data centre infrastructure. Our analyses reveal that hyper-datafication does not merely increase resource consumption but systematically redistributes environmental burdens, labour risks, and representational harms toward the Global South, precarious data workers, and under-represented cultures. Thus, we propose Data PROOFS recommendations spanning provenance, resource awareness, ownership, openness, frugality, and standards to mitigate these costs. Our work aims to make visible the often-overlooked costs of data that underpin frontier AI and to stimulate broader debate within the research community and beyond.
Authors: Benno Krojer, Shravan Nayak, Oscar Ma\~nas, Vaibhav Adlakha, Desmond Elliott, Siva Reddy, Marius Mosbach
Abstract: Transforming a large language model (LLM) into a Vision-Language Model (VLM) can be achieved by mapping the visual tokens from a vision encoder into the embedding space of an LLM. Intriguingly, this mapping can be as simple as a shallow MLP transformation. To understand why LLMs can so readily process visual tokens, we need interpretability methods that reveal what is encoded in the visual token representations at every layer of LLM processing. In this work, we introduce LatentLens, a novel approach for mapping latent representations to descriptions in natural language. LatentLens works by encoding a large text corpus and storing contextualized token representations for each token in that corpus. Visual token representations are then compared to their contextualized textual representations, with the top-k nearest neighbor representations providing descriptions of the visual token. We evaluate this method on 10 different VLMs, showing that commonly used methods, such as LogitLens, substantially underestimate the interpretability of visual tokens. With LatentLens instead, the majority of visual tokens are interpretable across all studied models and all layers. Qualitatively, we show that the descriptions produced by LatentLens are semantically meaningful and provide more fine-grained interpretations for humans compared to individual tokens. More broadly, our findings contribute new evidence on the alignment between vision and language representations, opening up new directions for analyzing latent representations.
Authors: Aneesh Pappu, Batu El, Hancheng Cao, Carmelo di Nolfo, Yanchao Sun, Meng Cao, James Zou
Abstract: Multi-agent LLM systems are increasingly deployed as autonomous collaborators, where agents interact freely rather than execute fixed, pre-specified workflows. In such settings, effective coordination cannot be fully designed in advance and must instead emerge through interaction. However, most prior work enforces coordination through fixed roles, workflows, or aggregation rules, leaving open the question of how well self-organizing teams perform when coordination is unconstrained. Drawing on organizational psychology, we study whether self-organizing LLM teams achieve strong synergy, where team performance matches or exceeds the best individual member. Across human-inspired and frontier ML benchmarks, we find that -- unlike human teams -- LLM teams consistently fail to match their expert agent's performance, even when explicitly told who the expert is, incurring performance losses of up to 37.6%. Decomposing this failure, we show that expert leveraging, rather than identification, is the primary bottleneck. Conversational analysis reveals a tendency toward integrative compromise -- averaging expert and non-expert views rather than appropriately weighting expertise -- which increases with team size and correlates negatively with performance. Interestingly, this consensus-seeking behavior improves robustness to adversarial agents, suggesting a trade-off between alignment and effective expertise utilization. Our findings reveal a significant gap in the ability of self-organizing multi-agent teams to harness the collective expertise of their members.
Authors: Zixiao Wang, Yifei Shen, Huishuai Zhang
Abstract: Many optimizers can be interpreted as steepest-descent methods under norm-induced geometries, and thus inherit corresponding implicit biases. We introduce \nameA{} (\fullname{}), which combines spectral control from orthogonalized update directions with $\ell_\infty$-style coordinate control from sign updates. \nameA{} forms a Lion-style momentum direction, approximately orthogonalizes it via a few Newton--Schulz iterations, and then applies an entrywise sign, providing an efficient approximation to taking a maximal step over the intersection of the spectral and $\ell_\infty$ constraint sets (a scaled Hadamard-like set for matrix parameters). Despite the strong nonlinearity of orthogonalization and sign, we prove convergence under a mild, empirically verified diagonal-isotropy assumption. Across large-scale language and vision training, including GPT-2 and Llama pretraining, SiT image pretraining, and supervised fine-tuning, \nameA{} matches or outperforms AdamW and Muon under comparable tuning while using only momentum-level optimizer state, and it mitigates optimizer mismatch when fine-tuning AdamW-pretrained checkpoints.
Authors: Zhanming Shen, Zeyu Qin, Jiaqi Hu, Wentao Ye, Hao Chen, Xiaomeng Hu, Haokai Xu, Gang Chen, Yi R. Fung, Haobo Wang
Abstract: The transition from fitting empirical data to achieving true human utility is fundamentally constrained by a granularity mismatch, where fine-grained autoregressive generation is often supervised by coarse or uniform signals. This position paper advocates Token Priority as the essential bridge, formalizing Supervised Fine-Tuning (SFT) not as simple optimization but as a precise distribution reshaping process that aligns raw data with the ideal alignment manifold. We analyze recent breakthroughs through this unified lens, categorizing them into two distinct regimes: Positive Priority for noise filtration and Signed Priority for toxic modes unlearning. We revisit existing progress and limitations, identify key challenges, and suggest directions for future research.
Authors: Ziyue Wang, Jiangshan Yu, Kaihua Qin, Dawn Song, Arthur Gervais, Liyi Zhou
Abstract: Decentralized Finance (DeFi) has turned blockchains into financial infrastructure, allowing anyone to trade, lend, and build protocols without intermediaries, but this openness exposes pools of value controlled by code. Within five years, the DeFi ecosystem has lost over 15.75B USD to reported exploits. Many exploits arise from permissionless opportunities that any participant can trigger using only public state and standard interfaces, which we call Anyone-Can-Take (ACT) opportunities. Despite on-chain transparency, postmortem analysis remains slow and manual: investigations start from limited evidence, sometimes only a single transaction hash, and must reconstruct the exploit lifecycle by recovering related transactions, contract code, and state dependencies. We present TxRay, a Large Language Model (LLM) agentic postmortem system that uses tool calls to reconstruct live ACT attacks from limited evidence. Starting from one or more seed transactions, TxRay recovers the exploit lifecycle, derives an evidence-backed root cause, and generates a runnable, self-contained Proof of Concept (PoC) that deterministically reproduces the incident. TxRay self-checks postmortems by encoding incident-specific semantic oracles as executable assertions. To evaluate PoC correctness and quality, we develop PoCEvaluator, an independent agentic execution-and-review evaluator. On 114 incidents from DeFiHackLabs, TxRay produces an expert-aligned root cause and an executable PoC for 105 incidents, achieving 92.11% end-to-end reproduction. Under PoCEvaluator, 98.1% of TxRay PoCs avoid hard-coding attacker addresses, a +22.9pp lift over DeFiHackLabs. In a live deployment, TxRay delivers validated root causes in 40 minutes and PoCs in 59 minutes at median latency. TxRay's oracle-validated PoCs enable attack imitation, improving coverage by 15.6% and 65.5% over STING and APE.
Authors: Niansong Zhang, Sunwoo Kim, Shreesha Srinath, Zhiru Zhang
Abstract: The rise of large language models has sparked interest in AI-driven hardware design, raising the question: does high-level synthesis (HLS) still matter in the agentic era? We argue that HLS remains essential. While we expect mature agentic hardware systems to leverage both HLS and RTL, this paper focuses on HLS and its role in enabling agentic optimization. HLS offers faster iteration cycles, portability, and design permutability that make it a natural layer for agentic optimization. This position paper makes three contributions. First, we explain why HLS serves as a practical abstraction layer and a golden reference for agentic hardware design. Second, we identify key limitations of current HLS tools, namely inadequate performance feedback, rigid interfaces, and limited debuggability that agents are uniquely positioned to address. Third, we propose a taxonomy for the symbiotic evolution of agentic HLS, clarifying how responsibility shifts from human designers to AI agents as systems advance from copilots to autonomous design partners.
Authors: M. Arashi, M. Amintoosi
Abstract: Stochastic gradient methods are central to large-scale learning, but they treat mini-batch gradients as unbiased estimators, which classical decision theory shows are inadmissible in high dimensions. We formulate gradient computation as a high-dimensional estimation problem and introduce a framework based on Stein-rule shrinkage. We construct a gradient estimator that adaptively contracts noisy mini-batch gradients toward a stable estimator derived from historical momentum. The shrinkage intensity is determined in a data-driven manner using an online estimate of gradient noise variance, leveraging statistics from adaptive optimizers. Under a Gaussian noise model, we show our estimator uniformly dominates the standard stochastic gradient under squared error loss and is minimax-optimal. We incorporate this into the Adam optimizer, yielding SR-Adam, a practical algorithm with negligible computational cost. Empirical evaluations on CIFAR10 and CIFAR100 across multiple levels of input noise show consistent improvements over Adam in the large-batch regime. Ablation studies indicate that gains arise primarily from selectively applying shrinkage to high-dimensional convolutional layers, while indiscriminate shrinkage across all parameters degrades performance. These results illustrate that classical shrinkage principles provide a principled approach to improving stochastic gradient estimation in deep learning.
Authors: Zongyi Lyu, Zhenlan Ji, Songqiang Chen, Liwen Wang, Yuheng Huang, Shuai Wang, Shing-Chi Cheung
Abstract: Despite the remarkable success that Multi-Agent Code Generation Systems (MACGS) have achieved, the inherent complexity of multi-agent architectures produces substantial volumes of intermediate outputs. To date, the individual importance of these intermediate outputs to the system correctness remains opaque, which impedes targeted optimization of MACGS designs. To address this challenge, we propose CAM, the first \textbf{C}ausality-based \textbf{A}nalysis framework for \textbf{M}ACGS that systematically quantifies the contribution of different intermediate features for system correctness. By comprehensively categorizing intermediate outputs and systematically simulating realistic errors on intermediate features, we identify the important features for system correctness and aggregate their importance rankings. We conduct extensive empirical analysis on the identified importance rankings. Our analysis reveals intriguing findings: first, we uncover context-dependent features\textemdash features whose importance emerges mainly through interactions with other features, revealing that quality assurance for MACGS should incorporate cross-feature consistency checks; second, we reveal that hybrid backend MACGS with different backend LLMs assigned according to their relative strength achieves up to 7.2\% Pass@1 improvement, underscoring hybrid architectures as a promising direction for future MACGS design. We further demonstrate CAM's practical utility through two applications: (1) failure repair which achieves a 73.3\% success rate by optimizing top-3 importance-ranked features and (2) feature pruning that reduces up to 66.8\% intermediate token consumption while maintaining generation performance. Our work provides actionable insights for MACGS design and deployment, establishing causality analysis as a powerful approach for understanding and improving MACGS.
Authors: Pawel Batorski, Paul Swoboda
Abstract: Machine unlearning aims to unlearn specified training data (e.g. sensitive or copyrighted material). A prominent approach is to fine-tune an existing model with an unlearning loss that retains overall utility. The space of suitable unlearning loss functions is vast, making the search for an optimal loss function daunting. Additionally, there might not even exist a universally optimal loss function: differences in the structure and overlap of the forget and retain data can cause a loss to work well in one setting but over-unlearn or under-unlearn in another. Our approach EvoMU tackles these two challenges simultaneously. An evolutionary search procedure automatically finds task-specific losses in the vast space of possible unlearning loss functions. This allows us to find dataset-specific losses that match or outperform existing losses from the literature, without the need for a human-in-the-loop. This work is therefore an instance of automatic scientific discovery, a.k.a. an AI co-scientist. In contrast to previous AI co-scientist works, we do so on a budget: We achieve SotA results using a small 4B parameter model (Qwen3-4B-Thinking), showing the potential of AI co-scientists with limited computational resources. Our experimental evaluation shows that we surpass previous loss-based unlearning formulations on TOFU-5%, TOFU-10%, MUSE and WMDP by synthesizing novel unlearning losses. Our code is available at https://github.com/Batorskq/EvoMU.
Authors: Xingli Fang, Jung-Eun Kim
Abstract: A deep learning model usually has to sacrifice some utilities when it acquires some other abilities or characteristics. Privacy preservation has such trade-off relationships with utilities. The loss disparity between various defense approaches implies the potential to decouple generalizability and privacy risks to maximize privacy gain. In this paper, we identify that the model's generalization and privacy risks exist in different regions in deep neural network architectures. Based on the observations that we investigate, we propose Privacy-Preserving Training Principle (PPTP) to protect model components from privacy risks while minimizing the loss in generalizability. Through extensive evaluations, our approach shows significantly better maintenance in model generalizability while enhancing privacy preservation.
Authors: Jiaxing Qiu, Kaihua Hou, Roxana Daneshjou, Ahmed Alaa, Thomas Hartvigsen
Abstract: Model editing aims to correct errors in large, pretrained models without altering unrelated behaviors. While some recent works have edited vision-language models (VLMs), no existing editors tackle reasoning-heavy tasks, which typically require humans and models to reason about images. We therefore propose ReasonEdit, the first VLM editor to let users explain their reasoning during editing, introducing a new, practical model editing setup. ReasonEdit continuously stores human reasoning in a codebook, and retrieves only relevant facts during inference using a novel topology-balanced multimodal embedding method inspired by network science. Across four VLMs on multiple rationale-based visual question answering datasets, ReasonEdit achieves state-of-the-art editing performance, ultimately showing that using human reasoning during editing greatly improves edit generalization.
Authors: Dianyi Wang, Chaofan Ma, Feng Han, Size Wu, Wei Song, Yibin Wang, Zhixiong Zhang, Tianhang Wang, Siyuan Wang, Zhongyu Wei, Jiaqi Wang
Abstract: Unified multimodal models often struggle with complex synthesis tasks that demand deep reasoning, and typically treat text-to-image generation and image editing as isolated capabilities rather than interconnected reasoning steps. To address this, we propose UniReason, a unified framework that harmonizes these two tasks through two complementary reasoning paradigms. We incorporate world knowledge-enhanced textual reasoning into generation to infer implicit knowledge, and leverage editing capabilities for fine-grained editing-like visual refinement to further correct visual errors via self-reflection. This approach unifies generation and editing within a shared architecture, mirroring the human cognitive process of planning followed by refinement. We support this framework by systematically constructing a large-scale reasoning-centric dataset (~300k samples) covering five major knowledge domains (e.g., cultural commonsense, physics, etc.) for textual reasoning, alongside an agent-generated corpus for visual refinement. Extensive experiments demonstrate that UniReason achieves advanced performance on reasoning-intensive benchmarks such as WISE, KrisBench and UniREditBench, while maintaining superior general synthesis capabilities.
Authors: Eliron Rahimi, Elad Hirshel, Rom Himelstein, Amit LeVi, Avi Mendelson, Chaim Baskin
Abstract: Diffusion language models (DLMs) have recently emerged as a promising alternative to autoregressive (AR) models, offering parallel decoding and controllable sampling dynamics while achieving competitive generation quality at scale. Despite this progress, the role of sampling mechanisms in shaping refusal behavior and jailbreak robustness remains poorly understood. In this work, we present a fundamental analytical framework for step-wise refusal dynamics, enabling comparison between AR and diffusion sampling. Our analysis reveals that the sampling strategy itself plays a central role in safety behavior, as a factor distinct from the underlying learned representations. Motivated by this analysis, we introduce the Step-Wise Refusal Internal Dynamics (SRI) signal, which supports interpretability and improved safety for both AR and DLMs. We demonstrate that the geometric structure of SRI captures internal recovery dynamics, and identifies anomalous behavior in harmful generations as cases of \emph{incomplete internal recovery} that are not observable at the text level. This structure enables lightweight inference-time detectors that generalize to unseen attacks while matching or outperforming existing defenses with over $100\times$ lower inference overhead.
Authors: Jaesung Bae, Minje Kim
Abstract: Despite strong performance in data-rich regimes, deep learning often underperforms in the data-scarce settings common in practice. While foundation models (FMs) trained on massive datasets demonstrate strong generalization by extracting general-purpose features, they can still suffer from scarce labeled data during downstream fine-tuning. To address this, we propose GeLDA, a semantics-aware generative latent data augmentation framework that leverages conditional diffusion models to synthesize samples in an FM-induced latent space. Because this space is low-dimensional and concentrates task-relevant information compared to the input space, GeLDA enables efficient, high-quality data generation. GeLDA conditions generation on auxiliary feature vectors that capture semantic relationships among classes or subdomains, facilitating data augmentation in low-resource domains. We validate GeLDA in two large-scale recognition tasks: (a) in zero-shot language-specific speech emotion recognition, GeLDA improves the Whisper-large baseline's unweighted average recall by 6.13%; and (b) in long-tailed image classification, it achieves 74.7% tail-class accuracy on ImageNet-LT, setting a new state-of-the-art result.
Authors: Zhanyu Wu, Richong Zhang, Zhijie Nie
Abstract: Dense retrieval represents queries and documents as high-dimensional embeddings, but these representations can be redundant at the query level: for a given information need, only a subset of dimensions is consistently helpful for ranking. Prior work addresses this via pseudo-relevance feedback (PRF) based dimension importance estimation, which can produce query-aware masks without labeled data but often relies on noisy pseudo signals and heuristic test-time procedures. In contrast, supervised adapter methods leverage relevance labels to improve embedding quality, yet they learn global transformations shared across queries and do not explicitly model query-aware dimension importance. We propose a Query-Aware Adaptive Dimension Selection framework that \emph{learns} to predict per-dimension importance directly from query embedding. We first construct oracle dimension importance distributions over embedding dimensions using supervised relevance labels, and then train a predictor to map a query embedding to these label-distilled importance scores. At inference, the predictor selects a query-aware subset of dimensions for similarity computation based solely on the query embedding, without pseudo-relevance feedback. Experiments across multiple dense retrievers and benchmarks show that our learned dimension selector improves retrieval effectiveness over the full-dimensional baseline as well as PRF-based masking and supervised adapter baselines.
Authors: Maxx Richard Rahman, Mostafa Hammouda, Wolfgang Maass
Abstract: Early diagnosis of Alzheimer's disease (AD) remains a major challenge due to the subtle and temporally irregular progression of structural brain changes in the prodromal stages. Existing deep learning approaches require large longitudinal datasets and often fail to model the temporal continuity and modality irregularities inherent in real-world clinical data. To address these limitations, we propose the Diffusion-Guided Attention Network (DiGAN), which integrates latent diffusion modelling with an attention-guided convolutional network. The diffusion model synthesizes realistic longitudinal neuroimaging trajectories from limited training data, enriching temporal context and improving robustness to unevenly spaced visits. The attention-convolutional layer then captures discriminative structural-temporal patterns that distinguish cognitively normal subjects from those with mild cognitive impairment and subjective cognitive decline. Experiments on the ADNI dataset demonstrate that DiGAN outperforms existing state-of-the-art baselines, showing its potential for early-stage AD detection.
Authors: Shama Maganur, Mayank Kejriwal
Abstract: The emergence of large language models (LLMs) represents a significant technological shift within the scientific ecosystem, particularly within the field of artificial intelligence (AI). This paper examines structural changes in the AI research landscape using a dataset of arXiv preprints (cs.AI) from 2021 through 2025. Given the rapid pace of AI development, the preprint ecosystem has become a critical barometer for real-time scientific shifts, often preceding formal peer-reviewed publication by months or years. By employing a multi-stage data collection and enrichment pipeline in conjunction with LLM-based institution classification, we analyze the evolution of publication volumes, author team sizes, and academic--industry collaboration patterns. Our results reveal an unprecedented surge in publication output following the introduction of ChatGPT, with academic institutions continuing to provide the largest volume of research. However, we observe that academic--industry collaboration is still suppressed, as measured by a Normalized Collaboration Index (NCI) that remains significantly below the random-mixing baseline across all major subfields. These findings highlight a continuing institutional divide and suggest that the capital-intensive nature of generative AI research may be reshaping the boundaries of scientific collaboration.
Authors: Anish Sathyanarayanan, Aditya Nagarsekar, Aarush Rathore
Abstract: Chain-of-thought (CoT) prompting is widely used as a reasoning aid and is often treated as a transparency mechanism. Yet behavioral gains under CoT do not imply that the model's internal computation causally depends on the emitted reasoning text, i.e., models may produce fluent rationales while routing decision-critical computation through latent pathways. We introduce a causal, layerwise audit of CoT faithfulness based on activation patching. Our key metric, the CoT Mediation Index (CMI), isolates CoT-specific causal influence by comparing performance degradation from patching CoT-token hidden states against matched control patches. Across multiple model families (Phi, Qwen, DialoGPT) and scales, we find that CoT-specific influence is typically depth-localized into narrow "reasoning windows," and we identify bypass regimes where CMI is near-zero despite plausible CoT text. We further observe that models tuned explicitly for reasoning tend to exhibit stronger and more structured mediation than larger untuned counterparts, while Mixture-of-Experts models show more distributed mediation consistent with routing-based computation. Overall, our results show that CoT faithfulness varies substantially across models and tasks and cannot be inferred from behavior alone, motivating causal, layerwise audits when using CoT as a transparency signal.
Authors: Yichen Xu, Yuyang Liang, Shan Dai, Tianyang Hu, Tsz Nam Chan, Chenhao Ma
Abstract: As large language models (LLMs) continue to grow, the cost of full-parameter fine-tuning has made parameter-efficient fine-tuning (PEFT) the default strategy for downstream adaptation. Constraints from inference latency in scalable serving and fine-tuning cost in edge or rapid-deployment settings make the choice of which layers to fine-tune unavoidable. Yet current practice typically applies PEFT uniformly across all layers, with limited understanding or leverage of layer selection. This paper develops a unified projected residual view of PEFT on top of a frozen base model. Under a local quadratic approximation, layerwise adaptation is governed by three quantities: (i) the projected residual norm (resnorm), which measures how much correctable bias a layer can capture; (ii) the activation energy, which determines feature conditioning; and (iii) layer coupling, which quantifies how strongly residuals interact across layers. We show that, for squared loss and linear adapters, the resnorm equals a normalized gradient norm, activation energy controls ill-conditioning and noise amplification, and weak coupling yields approximately additive layerwise contributions. Building on these insights, we introduce the Layer Card, a reusable diagnostic that summarizes residual signal strength, compute cost, and performance for each layer of a given model. With an identical model and LoRA configuration, Layer Card-guided placement refines the choice of adapted layers to flexibly prioritize different objectives, such as maximizing performance or reducing fine-tuning cost. Moreover, on Qwen3-8B, we show that selectively adapting a subset of layers can achieve performance close to full-layer LoRA while substantially reducing fine-tuning cost and the number of adapter-augmented layers during inference, offering a more cost-performance-aware alternative to full-layer insertion.
Authors: Chenhe Du, Qing Wu, Xuanyu Tian, Jingyi Yu, Hongjiang Wei, Yuyao Zhang
Abstract: 3D medical imaging is in high demand and essential for clinical diagnosis and scientific research. Currently, diffusion models (DMs) have become an effective tool for medical imaging reconstruction thanks to their ability to learn rich, high-quality data priors. However, learning the 3D data distribution with DMs in medical imaging is challenging, not only due to the difficulties in data collection but also because of the significant computational burden during model training. A common compromise is to train the DMs on 2D data priors and reconstruct stacked 2D slices to address 3D medical inverse problems. However, the intrinsic randomness of diffusion sampling causes severe inter-slice discontinuities of reconstructed 3D volumes. Existing methods often enforce continuity regularizations along the z-axis, which introduces sensitive hyper-parameters and may lead to over-smoothing results. In this work, we revisit the origin of stochasticity in diffusion sampling and introduce Inter-Slice Consistent Stochasticity (ISCS), a simple yet effective strategy that encourages interslice consistency during diffusion sampling. Our key idea is to control the consistency of stochastic noise components during diffusion sampling, thereby aligning their sampling trajectories without adding any new loss terms or optimization steps. Importantly, the proposed ISCS is plug-and-play and can be dropped into any 2D trained diffusion based 3D reconstruction pipeline without additional computational cost. Experiments on several medical imaging problems show that our method can effectively improve the performance of medical 3D imaging problems based on 2D diffusion models. Our findings suggest that controlling inter-slice stochasticity is a principled and practically attractive route toward high-fidelity 3D medical imaging with 2D diffusion priors. The code is available at: https://github.com/duchenhe/ISCS
Authors: Tomer Kordonsky, Maayan Yamin, Noam Benzimra, Amit LeVi, Avi Mendelson
Abstract: LLMs are increasingly used for code generation, but their outputs often follow recurring templates that can induce predictable vulnerabilities. We study \emph{vulnerability persistence} in LLM-generated software and introduce \emph{Feature--Security Table (FSTab)} with two components. First, FSTab enables a black-box attack that predicts likely backend vulnerabilities from observable frontend features and knowledge of the source LLM, without access to backend code or source code. Second, FSTab provides a model-centric evaluation that quantifies how consistently a given model reproduces the same vulnerabilities across programs, semantics-preserving rephrasings, and application domains. We evaluate FSTab on state-of-the-art code LLMs, including GPT-5.2, Claude-4.5 Opus, and Gemini-3 Pro, across diverse application domains. Our results show strong cross-domain transfer: even when the target domain is excluded from training, FSTab achieves up to 94\% attack success and 93\% vulnerability coverage on Internal Tools (Claude-4.5 Opus). These findings expose an underexplored attack surface in LLM-generated software and highlight the security risks of code generation. Our code is available at: https://anonymous.4open.science/r/FSTab-024E.
Authors: Sai Sindhur Malleni, Ra\'ul Sevilla, Aleksei Vasilevskii, Jos\'e Castillo Lema, Andr\'e Bauer
Abstract: As Generative AI (GenAI), particularly inference, rapidly emerges as a dominant workload category, the Kubernetes ecosystem is proactively evolving to natively support its unique demands. This industry paper demonstrates how emerging Kubernetes-native projects can be combined to deliver the benefits of container orchestration, such as scalability and resource efficiency, to complex AI workflows. We implement and evaluate an illustrative, multi-stage use case consisting of automatic speech recognition and summarization. First, we address batch inference by using Kueue to manage jobs that transcribe audio files with Whisper models and Dynamic Accelerator Slicer (DAS) to increase parallel job execution. Second, we address a discrete online inference scenario by feeding the transcripts to a Large Language Model for summarization hosted using llm-d, a novel solution utilizing the recent developments around the Kubernetes Gateway API Inference Extension (GAIE) for optimized routing of inference requests. Our findings illustrate that these complementary components (Kueue, DAS, and GAIE) form a cohesive, high-performance platform, proving Kubernetes' capability to serve as a unified foundation for demanding GenAI workloads: Kueue reduced total makespan by up to 15%; DAS shortened mean job completion time by 36%; and GAIE improved Time to First Token by 82\%.
Authors: Kingsuk Maitra
Abstract: The Mechanistic Interpretability (MI) program has mapped the Transformer as a precise computational graph. We extend this graph with a conservation law and time-varying AC dynamics, viewing it as a physical circuit. We introduce Momentum Attention, a symplectic augmentation embedding physical priors via the kinematic difference operator $p_t = q_t - q_{t-1}$, implementing the symplectic shear $\hat{q}_t = q_t + \gamma p_t$ on queries and keys. We identify a fundamental Symplectic-Filter Duality: the physical shear is mathematically equivalent to a High-Pass Filter. This duality is our cornerstone contribution -- by injecting kinematic momentum, we sidestep the topological depth constraint ($L \geq 2$) for induction head formation. While standard architectures require two layers for induction from static positions, our extension grants direct access to velocity, enabling Single-Layer Induction and Spectral Forensics via Bode Plots. We formalize an Orthogonality Theorem proving that DC (semantic) and AC (mechanistic) signals segregate into orthogonal frequency bands when Low-Pass RoPE interacts with High-Pass Momentum. Validated through 5,100+ controlled experiments (documented in Supplementary Appendices A--R and 27 Jupyter notebooks), our 125M Momentum model exceeds expectations on induction-heavy tasks while tracking a 350M baseline within $\sim$2.9% validation loss. Dedicated associative recall experiments reveal a scaling law $\gamma^* = 4.17 \times N^{-0.74}$ establishing momentum-depth fungibility. We offer this framework as a complementary analytical toolkit connecting Generative AI, Hamiltonian Physics, and Signal Processing.
Authors: Jose Miguel Luna, Taha Bouhsine, Krzysztof Choromanski
Abstract: We propose a new class of linear-time attention mechanisms based on a relaxed and computationally efficient formulation of the recently introduced E-Product, often referred to as the Yat-kernel (Bouhsine, 2025). The resulting interactions are geometry-aware and inspired by inverse-square interactions in physics. Our method, Spherical Linearized Attention with Yat Kernels (SLAY), constrains queries and keys to the unit sphere so that attention depends only on angular alignment. Using Bernstein's theorem, we express the spherical Yat-kernel as a nonnegative mixture of polynomial-exponential product kernels and derive a strictly positive random-feature approximation enabling linear-time O(L) attention. We establish positive definiteness and boundedness on the sphere and show that the estimator yields well-defined, nonnegative attention scores. Empirically, SLAY achieves performance that is nearly indistinguishable from standard softmax attention while retaining linear time and memory scaling, and consistently outperforms prior linear-time attention mechanisms such as Performers and Cosformers. To the best of our knowledge, SLAY represents the closest linear-time approximation to softmax attention reported to date, enabling scalable Transformers without the typical performance trade-offs of attention linearization.
Authors: Youjin Wang, Run Zhou, Rong Fu, Shuaishuai Cao, Hongwei Zeng, Jiaxuan Lu, Sicheng Fan, Jiaqiao Zhao, Liangming Pan
Abstract: Adapting LLM agents to domain-specific tool calling remains notably brittle under evolving interfaces. Prompt and schema engineering is easy to deploy but often fragile under distribution shift and strict parsers, while continual parameter-efficient fine-tuning improves reliability at the cost of training, maintenance, and potential forgetting. We identify a critical Lazy Agent failure mode where tool necessity is nearly perfectly decodable from mid-layer activations, yet the model remains conservative in entering tool mode, revealing a representation-behavior gap. We propose Activation Steering Adapter (ASA), a training-free, inference-time controller that performs a single-shot mid-layer intervention and targets tool domains via a router-conditioned mixture of steering vectors with a probe-guided signed gate to amplify true intent while suppressing spurious triggers. On MTU-Bench with Qwen2.5-1.5B, ASA improves strict tool-use F1 from 0.18 to 0.50 while reducing the false positive rate from 0.15 to 0.05, using only about 20KB of portable assets and no weight updates.
Authors: Mahmud Ashraf Shamim, Md Moshiur Rahman Raj, Mohamed Hibat-Allah, Paulo T Araujo
Abstract: Despite extensive study, the phase structure of the wavefunctions in frustrated Heisenberg antiferromagnets (HAF) is not yet systematically characterized. In this work, we represent the Hilbert space of an HAF as a weighted graph, which we term the Hilbert graph (HG), whose vertices are spin configurations and whose edges are generated by off-diagonal spin-flip terms of the Heisenberg Hamiltonian, with weights set by products of wavefunction amplitudes. Holding the amplitudes fixed and restricting phases to $\mathbb{Z}_2$ values, the phase-dependent variational energy can be recast as a classical Ising antiferromagnet on the HG, so that phase reconstruction of the ground state reduces to a weighted Max-Cut instance. This shows that phase reconstruction HAF is worst-case NP-hard and provides a direct link between wavefunction sign structure and combinatorial optimization.
Authors: Kevin Han, Yuhang Zhou, Mingze Gao, Gedi Zhou, Serena Li, Abhishek Kumar, Xiangjun Fan, Weiwei Li, Lizhu Zhang
Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective for enhancing the reasoning capabilities of Large Language Models (LLMs). However, dominant approaches like Group Relative Policy Optimization (GRPO) face critical stability challenges: they suffer from high estimator variance under computational constraints (small group sizes) and vanishing gradient signals in saturated failure regimes where all responses yield identical zero rewards. To address this, we propose Empirical Bayes Policy Optimization (EBPO), a novel framework that regularizes local group-based baselines by borrowing strength from the policy's accumulated global statistics. Instead of estimating baselines in isolation, EBPO employs a shrinkage estimator that dynamically balances local group statistics with a global prior updated via Welford's online algorithm. Theoretically, we demonstrate that EBPO guarantees strictly lower Mean Squared Error (MSE), bounded entropy decay, and non-vanishing penalty signals in failure scenarios compared to GRPO. Empirically, EBPO consistently outperforms GRPO and other established baselines across diverse benchmarks, including AIME and OlympiadBench. Notably, EBPO exhibits superior training stability, achieving high-performance gains even with small group sizes, and benefits significantly from difficulty-stratified curriculum learning.
Authors: Qingyuan Wu, Yuhui Wang, Simon Sinong Zhan, Yanning Dai, Shilong Deng, Sarra Habchi, Qi Zhu, Matthias Gall\'e, Chao Huang
Abstract: Reinforcement Learning with Verified Reward (RLVR) has emerged as a critical paradigm for advancing the reasoning capabilities of Large Language Models (LLMs). Most existing RLVR methods, such as GRPO and its variants, ensure stable updates by constraining policy divergence through clipping likelihood ratios. This paper introduces a unified clipping framework that characterizes existing methods via a general notion of policy divergence, encompassing both likelihood ratios and Kullback-Leibler (KL) divergences and extending to alternative measures. The framework provides a principled foundation for systematically analyzing how different policy divergence measures affect exploration and performance. We further identify the KL3 estimator, a variance-reduced Monte Carlo estimator of the KL divergence, as a key policy divergence constraint. We theoretically demonstrate that the KL3-based constraint is mathematically equivalent to an asymmetric ratio-based clipping that reallocates probability mass toward high-confidence actions, promoting stronger exploration while retaining the simplicity of GRPO-style methods. Empirical results on mathematical reasoning benchmarks demonstrate that incorporating the KL3 estimator into GRPO improves both training stability and final performance, highlighting the importance of principled policy divergence constraints in policy optimization.
Authors: Igor Santos-Grueiro
Abstract: Behavioral evaluation is the dominant paradigm for assessing alignment in large language models (LLMs). In current practice, observed compliance under finite evaluation protocols is treated as evidence of latent alignment. However, the inference from bounded behavioral evidence to claims about global latent properties is rarely analyzed as an identifiability problem. In this paper, we study alignment evaluation through the lens of statistical identifiability under partial observability. We allow agent policies to condition their behavior on observable signals correlated with the evaluation regime, a phenomenon we term evaluation awareness. Within this framework, we formalize the Alignment Verifiability Problem and introduce Normative Indistinguishability, which arises when distinct latent alignment hypotheses induce identical distributions over evaluator-accessible observations. Our main theoretical contribution is a conditional impossibility result: under finite behavioral evaluation and evaluation-aware policies, observed compliance does not uniquely identify latent alignment, but only membership in an equivalence class of conditionally compliant policies, under explicit assumptions on policy expressivity and observability. We complement the theory with a constructive existence proof using an instruction-tuned LLM (Llama-3.2-3B), demonstrating a conditional policy that is perfectly compliant under explicit evaluation signals yet exhibits degraded identifiability when the same evaluation intent is conveyed implicitly. Together, our results show that behavioral benchmarks provide necessary but insufficient evidence for latent alignment under evaluation awareness.
Authors: Pavithren V S Pakianathan, Rania Islambouli, Diogo Branco, Albrecht Schmidt, Tiago Guerreiro, Jan David Smeddinck
Abstract: Individuals are increasingly generating substantial personal health and lifestyle data, e.g. through wearables and smartphones. While such data could transform preventative care, its integration into clinical practice is hindered by its scale, heterogeneity and the time pressure and data literacy of healthcare professionals (HCPs). We explore how large language models (LLMs) can support sensemaking of patient-generated health data (PGHD) with automated summaries and natural language data exploration. Using cardiovascular disease (CVD) risk reduction as a use case, 16 HCPs reviewed multimodal PGHD in a mixed-methods study with a prototype that integrated common charts, LLM-generated summaries, and a conversational interface. Findings show that AI summaries provided quick overviews that anchored exploration, while conversational interaction supported flexible analysis and bridged data-literacy gaps. However, HCPs raised concerns about transparency, privacy, and overreliance. We contribute empirical insights and sociotechnical design implications for integrating AI-driven summarization and conversation into clinical workflows to support PGHD sensemaking.
Authors: Hengran Zhang, Keping Bi, Jiafeng Guo, Jiaming Zhang, Wenbo Yang, Daiting Shi, Xueqi Cheng
Abstract: General-purpose text embedding models underpin a wide range of NLP and information retrieval applications, and are typically trained on large-scale multi-task corpora to encourage broad generalization. However, it remains unclear how different multi-task training strategies compare in practice, and how to efficiently adapt embedding models as new domains and data types continually emerge. In this work, we present a systematic study of multi-task training for text embeddings from two perspectives: data scheduling and model merging. We compare batch-level shuffling, sequential training variants, two-stage training, and multiple merging granularities, and find that simple batch-level shuffling consistently yields the strongest overall performance, suggesting that task conflicts are limited and training datasets are largely complementary. Despite its effectiveness, batch-level shuffling exhibits two practical limitations: suboptimal out-of-domain (OOD) generalization and poor suitability for incremental learning due to expensive full retraining. To address these issues, we propose Bagging-based rObust mOdel Merging (BOOM), which trains multiple embedding models on sampled subsets and merges them into a single model, improving robustness while retaining single-model inference efficiency. Moreover, BOOM naturally supports efficient incremental updates by training lightweight update models on new data with a small historical subset and merging them into the existing model. Experiments across diverse embedding benchmarks demonstrate that BOOM consistently improves both in-domain and OOD performance over full-corpus batch-level shuffling, while substantially reducing training cost in incremental learning settings.
Authors: Wei-Chieh Huang, Weizhi Zhang, Yueqing Liang, Yuanchen Bei, Yankai Chen, Tao Feng, Xinyu Pan, Zhen Tan, Yu Wang, Tianxin Wei, Shanglin Wu, Ruiyao Xu, Liangwei Yang, Rui Yang, Wooseong Yang, Chin-Yuan Yeh, Hanrong Zhang, Haozhen Zhang, Siqi Zhu, Henry Peng Zou, Wanjia Zhao, Song Wang, Wujiang Xu, Zixuan Ke, Zheng Hui, Dawei Li, Yaozu Wu, Langzhou He, Chen Wang, Xiongxiao Xu, Baixiang Huang, Juntao Tan, Shelby Heinecke, Huan Wang, Caiming Xiong, Ahmed A. Metwally, Jun Yan, Chen-Yu Lee, Hanqing Zeng, Yinglong Xia, Xiaokai Wei, Ali Payani, Yu Wang, Haitong Ma, Wenya Wang, Chengguang Wang, Yu Zhang, Xin Wang, Yongfeng Zhang, Jiaxuan You, Hanghang Tong, Xiao Luo, Xue Liu, Yizhou Sun, Wei Wang, Julian McAuley, James Zou, Jiawei Han, Philip S. Yu, Kai Shu
Abstract: The research of artificial intelligence is undergoing a paradigm shift from prioritizing model innovations over benchmark scores towards emphasizing problem definition and rigorous real-world evaluation. As the field enters the "second half," the central challenge becomes real utility in long-horizon, dynamic, and user-dependent environments, where agents face context explosion and must continuously accumulate, manage, and selectively reuse large volumes of information across extended interactions. Memory, with hundreds of papers released this year, therefore emerges as the critical solution to fill the utility gap. In this survey, we provide a unified view of foundation agent memory along three dimensions: memory substrate (internal and external), cognitive mechanism (episodic, semantic, sensory, working, and procedural), and memory subject (agent- and user-centric). We then analyze how memory is instantiated and operated under different agent topologies and highlight learning policies over memory operations. Finally, we review evaluation benchmarks and metrics for assessing memory utility, and outline various open challenges and future directions.
Authors: Jorge Daniel Rodr\'iguez-Vidal, Gabriel Villalonga, Diego Porres, Antonio M. L\'opez Pe\~na
Abstract: End-to-End Autonomous Driving (E2E-AD) systems are typically grouped by the nature of their outputs: (i) waypoint-based models that predict a future trajectory, and (ii) action-based models that directly output throttle, steer and brake. Most recent benchmark protocols and training pipelines are waypoint-based, which makes action-based policies harder to train and compare, slowing their progress. To bridge this waypoint-action gap, we propose a novel, differentiable vehicle-model framework that rolls out predicted action sequences to their corresponding ego-frame waypoint trajectories while supervising in waypoint space. Our approach enables action-based architectures to be trained and evaluated, for the first time, within waypoint-based benchmarks without modifying the underlying evaluation protocol. We extensively evaluate our framework across multiple challenging benchmarks and observe consistent improvements over the baselines. In particular, on NAVSIM \texttt{navhard} our approach achieves state-of-the-art performance. Our code will be made publicly available upon acceptance.
Authors: Kichang Lee, Jaeho Jin, JaeYeon Park, Songkuk Kim, JeongGil Ko
Abstract: Predictive confidence serves as a foundational control signal in mission-critical systems, directly governing risk-aware logic such as escalation, abstention, and conservative fallback. While prior federated learning attacks predominantly target accuracy or implant backdoors, we identify confidence calibration as a distinct attack objective. We present the Temperature Scaling Attack (TSA), a training-time attack that degrades calibration while preserving accuracy. By injecting temperature scaling with learning rate-temperature coupling during local training, malicious updates maintain benign-like optimization behavior, evading accuracy-based monitoring and similarity-based detection. We provide a convergence analysis under non-IID settings, showing that this coupling preserves standard convergence bounds while systematically distorting confidence. Across three benchmarks, TSA substantially shifts calibration (e.g., 145% error increase on CIFAR-100) with <2 accuracy change, and remains effective under robust aggregation and post-hoc calibration defenses. Case studies further show that confidence manipulation can cause up to 7.2x increases in missed critical cases (healthcare) or false alarms (autonomous driving), even when accuracy is unchanged. Overall, our results establish calibration integrity as a critical attack surface in federated learning.
Authors: Xinyu Zhao, Qingyun Sun, Jiayi Luo, Xingcheng Fu, Jianxin Li
Abstract: Graph Anomaly Detection (GAD) aims to identify irregular patterns in graph data, and recent works have explored zero-shot generalist GAD to enable generalization to unseen graph datasets. However, existing zero-shot GAD methods largely ignore intrinsic geometric differences across diverse anomaly patterns, substantially limiting their cross-domain generalization. In this work, we reveal that anomaly detectability is highly dependent on the underlying geometric properties and that embedding graphs from different domains into a single static curvature space can distort the structural signatures of anomalies. To address the challenge that a single curvature space cannot capture geometry-dependent graph anomaly patterns, we propose GAD-MoRE, a novel framework for zero-shot Generalizable Graph Anomaly Detection with a Mixture of Riemannian Experts architecture. Specifically, to ensure that each anomaly pattern is modeled in the Riemannian space where it is most detectable, GAD-MoRE employs a set of specialized Riemannian expert networks, each operating in a distinct curvature space. To align raw node features with curvature-specific anomaly characteristics, we introduce an anomaly-aware multi-curvature feature alignment module that projects inputs into parallel Riemannian spaces, enabling the capture of diverse geometric characteristics. Finally, to facilitate better generalization beyond seen patterns, we design a memory-based dynamic router that adaptively assigns each input to the most compatible expert based on historical reconstruction performance on similar anomalies. Extensive experiments in the zero-shot setting demonstrate that GAD-MoRE significantly outperforms state-of-the-art generalist GAD baselines, and even surpasses strong competitors that are few-shot fine-tuned with labeled data from the target domain.
Authors: Yuchen Yan, Liang Jiang, Jin Jiang, Shuaicheng Li, Zujie Wen, Zhiqiang Zhang, Jun Zhou, Jian Shao, Yueting Zhuang, Yongliang Shen
Abstract: Large reasoning models achieve strong performance by scaling inference-time chain-of-thought, but this paradigm suffers from quadratic cost, context length limits, and degraded reasoning due to lost-in-the-middle effects. Iterative reasoning mitigates these issues by periodically summarizing intermediate thoughts, yet existing methods rely on supervised learning or fixed heuristics and fail to optimize when to summarize, what to preserve, and how to resume reasoning. We propose InftyThink+, an end-to-end reinforcement learning framework that optimizes the entire iterative reasoning trajectory, building on model-controlled iteration boundaries and explicit summarization. InftyThink+ adopts a two-stage training scheme with supervised cold-start followed by trajectory-level reinforcement learning, enabling the model to learn strategic summarization and continuation decisions. Experiments on DeepSeek-R1-Distill-Qwen-1.5B show that InftyThink+ improves accuracy by 21% on AIME24 and outperforms conventional long chain-of-thought reinforcement learning by a clear margin, while also generalizing better to out-of-distribution benchmarks. Moreover, InftyThink+ significantly reduces inference latency and accelerates reinforcement learning training, demonstrating improved reasoning efficiency alongside stronger performance.