Authors: Jean-Marc Fellous, Gert Cauwenberghs, Cornelia Ferm\"uller, Yulia Sandamisrkaya, Terrence Sejnowski
Abstract: Neuroscience and Artificial Intelligence (AI) have made significant progress in the past few years but have only been loosely inter-connected. Based on a workshop held in August 2025, we identify current and future areas of synergism between these two fields. We focus on the subareas of embodiment, language and communication, robotics, learning in humans and machines and Neuromorphic engineering to take stock of the progress made so far, and possible promising new future avenues. Overall, we advocate for the development of NeuroAI, a type of Neuroscience-informed Artificial Intelligence that, we argue, has the potential for significantly improving the scope and efficiency of AI algorithms while simultaneously changing the way we understand biological neural computations. We include personal statements from several leading researchers on their diverse views of NeuroAI. Two Strength-Weakness-Opportunities-Threat (SWOT) analyses by researchers and trainees are appended that describe the benefits and risks offered by NeuroAI.
Authors: Shuhui Qu
Abstract: Inference-time planning with large language models frequently breaks under partial observability: when task-critical preconditions are not specified at query time, models tend to hallucinate missing facts or produce plans that violate hard constraints. We introduce \textbf{Self-Querying Bidirectional Categorical Planning (SQ-BCP)}, which explicitly represents precondition status (\texttt{Sat}/\texttt{Viol}/\texttt{Unk}) and resolves unknowns via (i) targeted self-queries to an oracle/user or (ii) \emph{bridging} hypotheses that establish the missing condition through an additional action. SQ-BCP performs bidirectional search and invokes a pullback-based verifier as a categorical certificate of goal compatibility, while using distance-based scores only for ranking and pruning. We prove that when the verifier succeeds and hard constraints pass deterministic checks, accepted plans are compatible with goal requirements; under bounded branching and finite resolution depth, SQ-BCP finds an accepting plan when one exists. Across WikiHow and RecipeNLG tasks with withheld preconditions, SQ-BCP reduces resource-violation rates to \textbf{14.9\%} and \textbf{5.8\%} (vs.\ \textbf{26.0\%} and \textbf{15.7\%} for the best baseline), while maintaining competitive reference quality.
Authors: Shuhui Qu
Abstract: Natural-language planning often involves vague predicates (e.g., suitable substitute, stable enough) whose satisfaction is inherently graded. Existing category-theoretic planners provide compositional structure and pullback-based hard-constraint verification, but treat applicability as crisp, forcing thresholding that collapses meaningful distinctions and cannot track quality degradation across multi-step plans. We propose Fuzzy Category-theoretic Planning (FCP), which annotates each action (morphism) with a degree in [0,1], composes plan quality via a t-norm Lukasiewicz, and retains crisp executability checks via pullback verification. FCP grounds graded applicability from language using an LLM with k-sample median aggregation and supports meeting-in-the-middle search using residuum-based backward requirements. We evaluate on (i) public PDDL3 preference/oversubscription benchmarks and (ii) RecipeNLG-Subs, a missing-substitute recipe-planning benchmark built from RecipeNLG with substitution candidates from Recipe1MSubs and FoodKG. FCP improves success and reduces hard-constraint violations on RecipeNLG-Subs compared to LLM-only and ReAct-style baselines, while remaining competitive with classical PDDL3 planners.
Authors: Jincheng Bai, Zhenyu Zhang, Jennifer Zhang, Zhihuai Zhu
Abstract: Today, E-commerce sellers face several key challenges, including difficulties in discovering and effectively utilizing available programs and tools, and struggling to understand and utilize rich data from various tools. We therefore aim to develop Insight Agents (IA), a conversational multi-agent Data Insight system, to provide E-commerce sellers with personalized data and business insights through automated information retrieval. Our hypothesis is that IA will serve as a force multiplier for sellers, thereby driving incremental seller adoption by reducing the effort required and increase speed at which sellers make good business decisions. In this paper, we introduce this novel LLM-backed end-to-end agentic system built on a plan-and-execute paradigm and designed for comprehensive coverage, high accuracy, and low latency. It features a hierarchical multi-agent structure, consisting of manager agent and two worker agents: data presentation and insight generation, for efficient information retrieval and problem-solving. We design a simple yet effective ML solution for manager agent that combines Out-of-Domain (OOD) detection using a lightweight encoder-decoder model and agent routing through a BERT-based classifier, optimizing both accuracy and latency. Within the two worker agents, a strategic planning is designed for API-based data model that breaks down queries into granular components to generate more accurate responses, and domain knowledge is dynamically injected to to enhance the insight generator. IA has been launched for Amazon sellers in US, which has achieved high accuracy of 90% based on human evaluation, with latency of P90 below 15s.
Authors: Amirmohammad Farzaneh, Salvatore D'Oro, Osvaldo Simeone
Abstract: Large language model (LLM)-powered agents can translate high-level user intents into plans and actions in an environment. Yet after observing an outcome, users may wonder: What if I had phrased my intent differently? We introduce a framework that enables such counterfactual reasoning in agentic LLM-driven control scenarios, while providing formal reliability guarantees. Our approach models the closed-loop interaction between a user, an LLM-based agent, and an environment as a structural causal model (SCM), and leverages test-time scaling to generate multiple candidate counterfactual outcomes via probabilistic abduction. Through an offline calibration phase, the proposed conformal counterfactual generation (CCG) yields sets of counterfactual outcomes that are guaranteed to contain the true counterfactual outcome with high probability. We showcase the performance of CCG on a wireless network control use case, demonstrating significant advantages compared to naive re-execution baselines.
Authors: Zixuan Xiao, Chunguang Hu, Jun Ma
Abstract: As an important part of urbanization, the development monitoring of newly constructed parks is of great significance for evaluating the effect of urban planning and optimizing resource allocation. However, traditional change detection methods based on remote sensing imagery have obvious limitations in high-level and intelligent analysis, and thus are difficult to meet the requirements of current urban planning and management. In face of the growing demand for complex multi-modal data analysis in urban park development monitoring, these methods often fail to provide flexible analysis capabilities for diverse application scenarios. This study proposes a multi-modal LLM agent framework, which aims to make full use of the semantic understanding and reasoning capabilities of LLM to meet the challenges in urban park development monitoring. In this framework, a general horizontal and vertical data alignment mechanism is designed to ensure the consistency and effective tracking of multi-modal data. At the same time, a specific toolkit is constructed to alleviate the hallucination issues of LLM due to the lack of domain-specific knowledge. Compared to vanilla GPT-4o and other agents, our approach enables robust multi-modal information fusion and analysis, offering reliable and scalable solutions tailored to the diverse and evolving demands of urban park development monitoring.
Authors: Hang Zhang, Ruheng Wang, Yuelyu Ji, Mingu Kwak, Xizhi Wu, Chenyu Li, Li Zhang, Wenqi Shi, Yifan Peng, Yanshan Wang
Abstract: Large language models have achieved strong performance on medical reasoning benchmarks, yet their deployment in clinical settings demands rigorous verification to ensure factual accuracy. While reward models offer a scalable approach for reasoning trace verification, existing methods face two limitations: they produce only scalar reward values without explicit justification, and they rely on single-pass retrieval that precludes adaptive knowledge access as verification unfolds. We introduce $\method$, an agentic framework that addresses these limitations by training medical reasoning verifiers to iteratively query external medical corpora during evaluation. Our approach combines tool-augmented verification with an iterative reinforcement learning paradigm that requires only trace-level supervision, alongside an adaptive curriculum mechanism that dynamically adjusts training data distribution. Across four medical reasoning benchmarks, $\method$ achieves substantial gains over existing methods, improving MedQA accuracy by 23.5% and MedXpertQA by 32.0% relative to the base generator in particular. Crucially, $\method$ demonstrates an $\mathbf{8\times}$ reduction in sampling budget requirement compared to prior reward model baselines. These findings establish that grounding verification in dynamically retrieved evidence offers a principled path toward more reliable medical reasoning systems.
Authors: Zhenchen Tang, Songlin Yang, Zichuan Wang, Bo Peng, Yang Li, Beibei Dong, Jing Dong
Abstract: Unified Multimodal Models (UMMs) exhibit strong understanding, yet this capability often fails to effectively guide generation. We identify this as a Cognitive Gap: the model lacks the understanding of how to enhance its own generation process. To bridge this gap, we propose Endogenous Reprompting, a mechanism that transforms the model's understanding from a passive encoding process into an explicit generative reasoning step by generating self-aligned descriptors during generation. To achieve this, we introduce SEER (Self-Evolving Evaluator and Reprompter), a training framework that establishes a two-stage endogenous loop using only 300 samples from a compact proxy task, Visual Instruction Elaboration. First, Reinforcement Learning with Verifiable Rewards (RLVR) activates the model's latent evaluation ability via curriculum learning, producing a high-fidelity endogenous reward signal. Second, Reinforcement Learning with Model-rewarded Thinking (RLMT) leverages this signal to optimize the generative reasoning policy. Experiments show that SEER consistently outperforms state-of-the-art baselines in evaluation accuracy, reprompting efficiency, and generation quality, without sacrificing general multimodal capabilities.
Authors: Hyunseung Chung, Jungwoo Oh, Daeun Kyung, Jiho Kim, Yeonsu Kwon, Min-Gyu Kim, Edward Choi
Abstract: Recent advances in Multimodal Large Language Models have rapidly expanded to electrocardiograms, focusing on classification, report generation, and single-turn QA tasks. However, these models fall short in real-world scenarios, lacking multi-turn conversational ability, on-device efficiency, and precise understanding of ECG measurements such as the PQRST intervals. To address these limitations, we introduce ECG-Agent, the first LLM-based tool-calling agent for multi-turn ECG dialogue. To facilitate its development and evaluation, we also present ECG-Multi-Turn-Dialogue (ECG-MTD) dataset, a collection of realistic user-assistant multi-turn dialogues for diverse ECG lead configurations. We develop ECG-Agents in various sizes, from on-device capable to larger agents. Experimental results show that ECG-Agents outperform baseline ECG-LLMs in response accuracy. Furthermore, on-device agents achieve comparable performance to larger agents in various evaluations that assess response accuracy, tool-calling ability, and hallucinations, demonstrating their viability for real-world applications.
Authors: Weiquan Huang, Zixuan Wang, Hehai Lin, Sudong Wang, Bo Xu, Qian Li, Beier Zhu, Linyi Yang, Chengwei Qin
Abstract: The rapid evolution of Large Language Model (LLM) agents has necessitated robust memory systems to support cohesive long-term interaction and complex reasoning. Benefiting from the strong capabilities of LLMs, recent research focus has shifted from simple context extension to the development of dedicated agentic memory systems. However, existing approaches typically rely on rigid retrieval granularity, accumulation-heavy maintenance strategies, and coarse-grained update mechanisms. These design choices create a persistent mismatch between stored information and task-specific reasoning demands, while leading to the unchecked accumulation of logical inconsistencies over time. To address these challenges, we propose Adaptive Memory via Multi-Agent Collaboration (AMA), a novel framework that leverages coordinated agents to manage memory across multiple granularities. AMA employs a hierarchical memory design that dynamically aligns retrieval granularity with task complexity. Specifically, the Constructor and Retriever jointly enable multi-granularity memory construction and adaptive query routing. The Judge verifies the relevance and consistency of retrieved content, triggering iterative retrieval when evidence is insufficient or invoking the Refresher upon detecting logical conflicts. The Refresher then enforces memory consistency by performing targeted updates or removing outdated entries. Extensive experiments on challenging long-context benchmarks show that AMA significantly outperforms state-of-the-art baselines while reducing token consumption by approximately 80% compared to full-context methods, demonstrating its effectiveness in maintaining retrieval precision and long-term memory consistency.
Authors: Zhengbo Jiao, Hongyu Xian, Qinglong Wang, Yunpu Ma, Zhebo Wang, Zifan Zhang, Dezhang Kong, Meng Han
Abstract: Large language models (LLMs) struggle with complex, long-horizon reasoning due to instability caused by their frozen policy assumption. Current test-time scaling methods treat execution feedback merely as an external signal for filtering or rewriting trajectories, without internalizing it to improve the underlying reasoning strategy. Inspired by Popper's epistemology of "conjectures and refutations," we argue that intelligence requires real-time evolution of the model's policy through learning from failed attempts. We introduce Policy of Thoughts (PoT), a framework that recasts reasoning as a within-instance online optimization process. PoT first generates diverse candidate solutions via an efficient exploration mechanism, then uses Group Relative Policy Optimization (GRPO) to update a transient LoRA adapter based on execution feedback. This closed-loop design enables dynamic, instance-specific refinement of the model's reasoning priors. Experiments show that PoT dramatically boosts performance: a 4B model achieves 49.71% accuracy on LiveCodeBench, outperforming GPT-4o and DeepSeek-V3 despite being over 50 smaller.
Authors: Le Zhang, Yixiong Xiao, Xinjiang Lu, Jingjia Cao, Yusai Zhao, Jingbo Zhou, Lang An, Zikan Feng, Wanxiang Sha, Yu Shi, Congxi Xiao, Jian Xiong, Yankai Zhang, Hua Wu, Haifeng Wang
Abstract: Graphical User Interface (GUI) agents show great potential for enabling foundation models to complete real-world tasks, revolutionizing human-computer interaction and improving human productivity. In this report, we present OmegaUse, a general-purpose GUI agent model for autonomous task execution on both mobile and desktop platforms, supporting computer-use and phone-use scenarios. Building an effective GUI agent model relies on two factors: (1) high-quality data and (2) effective training methods. To address these, we introduce a carefully engineered data-construction pipeline and a decoupled training paradigm. For data construction, we leverage rigorously curated open-source datasets and introduce a novel automated synthesis framework that integrates bottom-up autonomous exploration with top-down taxonomy-guided generation to create high-fidelity synthetic data. For training, to better leverage these data, we adopt a two-stage strategy: Supervised Fine-Tuning (SFT) to establish fundamental interaction syntax, followed by Group Relative Policy Optimization (GRPO) to improve spatial grounding and sequential planning. To balance computational efficiency with agentic reasoning capacity, OmegaUse is built on a Mixture-of-Experts (MoE) backbone. To evaluate cross-terminal capabilities in an offline setting, we introduce OS-Nav, a benchmark suite spanning multiple operating systems: ChiM-Nav, targeting Chinese Android mobile environments, and Ubu-Nav, focusing on routine desktop interactions on Ubuntu. Extensive experiments show that OmegaUse is highly competitive across established GUI benchmarks, achieving a state-of-the-art (SOTA) score of 96.3% on ScreenSpot-V2 and a leading 79.1% step success rate on AndroidControl. OmegaUse also performs strongly on OS-Nav, reaching 74.24% step success on ChiM-Nav and 55.9% average success on Ubu-Nav.
Authors: Zhenxuan Fan, Jie Cao, Yang Dai, Zheqi Lv, Wenqiao Zhang, Zhongle Xie, Peng LU, Beng Chin Ooi
Abstract: Chain-of-thought (CoT) prompting improves LLM reasoning but incurs high latency and memory cost due to verbose traces, motivating CoT compression with preserved correctness. Existing methods either shorten CoTs at the semantic level, which is often conservative, or prune tokens aggressively, which can miss task-critical cues and degrade accuracy. Moreover, combining the two is non-trivial due to sequential dependency, task-agnostic pruning, and distribution mismatch. We propose \textbf{CtrlCoT}, a dual-granularity CoT compression framework that harmonizes semantic abstraction and token-level pruning through three components: Hierarchical Reasoning Abstraction produces CoTs at multiple semantic granularities; Logic-Preserving Distillation trains a logic-aware pruner to retain indispensable reasoning cues (e.g., numbers and operators) across pruning ratios; and Distribution-Alignment Generation aligns compressed traces with fluent inference-time reasoning styles to avoid fragmentation. On MATH-500 with Qwen2.5-7B-Instruct, CtrlCoT uses 30.7\% fewer tokens while achieving 7.6 percentage points higher than the strongest baseline, demonstrating more efficient and reliable reasoning. Our code will be publicly available at https://github.com/fanzhenxuan/Ctrl-CoT.
Authors: Nico Mutzner, Taha Yasseri, Heiko Rauhut
Abstract: The introduction of artificial intelligence (AI) agents into human group settings raises essential questions about how these novel participants influence cooperative social norms. While previous studies on human-AI cooperation have primarily focused on dyadic interactions, little is known about how integrating AI agents affects the emergence and maintenance of cooperative norms in small groups. This study addresses this gap through an online experiment using a repeated four-player Public Goods Game (PGG). Each group consisted of three human participants and one bot, which was framed either as human or AI and followed one of three predefined decision strategies: unconditional cooperation, conditional cooperation, or free-riding. In our sample of 236 participants, we found that reciprocal group dynamics and behavioural inertia primarily drove cooperation. These normative mechanisms operated identically across conditions, resulting in cooperation levels that did not differ significantly between human and AI labels. Furthermore, we found no evidence of differences in norm persistence in a follow-up Prisoner's Dilemma, or in participants' normative perceptions. Participants' behaviour followed the same normative logic across human and AI conditions, indicating that cooperation depended on group behaviour rather than partner identity. This supports a pattern of normative equivalence, in which the mechanisms that sustain cooperation function similarly in mixed human-AI and all human groups. These findings suggest that cooperative norms are flexible enough to extend to artificial agents, blurring the boundary between humans and AI in collective decision-making.
Authors: Oguzhan Gungordu, Siheng Xiong, Faramarz Fekri
Abstract: Large Language Models (LLMs) have enabled automated heuristic design (AHD) for combinatorial optimization problems (COPs), but existing frameworks' reliance on fixed evolutionary rules and static prompt templates often leads to myopic heuristic generation, redundant evaluations, and limited reasoning about how new heuristics should be derived. We propose a novel multi-agent reasoning framework, referred to as Planning through World Model for Automated Heuristic Design via Self-Evolving LLMs (PathWise), which formulates heuristic generation as a sequential decision process over an entailment graph serving as a compact, stateful memory of the search trajectory. This approach allows the system to carry forward past decisions and reuse or avoid derivation information across generations. A policy agent plans evolutionary actions, a world model agent generates heuristic rollouts conditioned on those actions, and critic agents provide routed reflections summarizing lessons from prior steps, shifting LLM-based AHD from trial-and-error evolution toward state-aware planning through reasoning. Experiments across diverse COPs show that PathWise converges faster to better heuristics, generalizes across different LLM backbones, and scales to larger problem sizes.
Authors: Yaacov Pariente, Vadim Indelman
Abstract: We study risk-sensitive planning under partial observability using the dynamic risk measure Iterated Conditional Value-at-Risk (ICVaR). A policy evaluation algorithm for ICVaR is developed with finite-time performance guarantees that do not depend on the cardinality of the action space. Building on this foundation, three widely used online planning algorithms--Sparse Sampling, Particle Filter Trees with Double Progressive Widening (PFT-DPW), and Partially Observable Monte Carlo Planning with Observation Widening (POMCPOW)--are extended to optimize the ICVaR value function rather than the expectation of the return. Our formulations introduce a risk parameter $\alpha$, where $\alpha = 1$ recovers standard expectation-based planning and $\alpha < 1$ induces increasing risk aversion. For ICVaR Sparse Sampling, we establish finite-time performance guarantees under the risk-sensitive objective, which further enable a novel exploration strategy tailored to ICVaR. Experiments on benchmark POMDP domains demonstrate that the proposed ICVaR planners achieve lower tail risk compared to their risk-neutral counterparts.
Authors: Gray Cox
Abstract: This paper introduces a methodological framework for empirically testing AI alignment strategies through structured multi-model dialogue. Drawing on Peace Studies traditions - particularly interest-based negotiation, conflict transformation, and commons governance - we operationalize Viral Collaborative Wisdom (VCW), an approach that reframes alignment from a control problem to a relationship problem developed through dialogical reasoning. Our experimental design assigns four distinct roles (Proposer, Responder, Monitor, Translator) to different AI systems across six conditions, testing whether current large language models can engage substantively with complex alignment frameworks. Using Claude, Gemini, and GPT-4o, we conducted 72 dialogue turns totaling 576,822 characters of structured exchange. Results demonstrate that AI systems can engage meaningfully with Peace Studies concepts, surface complementary objections from different architectural perspectives, and generate emergent insights not present in initial framings - including the novel synthesis of "VCW as transitional framework." Cross-architecture patterns reveal that different models foreground different concerns: Claude emphasized verification challenges, Gemini focused on bias and scalability, and GPT-4o highlighted implementation barriers. The framework provides researchers with replicable methods for stress-testing alignment proposals before implementation, while the findings offer preliminary evidence about AI capacity for the kind of dialogical reasoning VCW proposes. We discuss limitations, including the observation that dialogues engaged more with process elements than with foundational claims about AI nature, and outline directions for future research including human-AI hybrid protocols and extended dialogue studies.
Authors: Yanqi Dai, Yuxiang Ji, Xiao Zhang, Yong Wang, Xiangxiang Chu, Zhiwu Lu
Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) offers a robust mechanism for enhancing mathematical reasoning in large models. However, we identify a systematic lack of emphasis on more challenging questions in existing methods from both algorithmic and data perspectives, despite their importance for refining underdeveloped capabilities. Algorithmically, widely used Group Relative Policy Optimization (GRPO) suffers from an implicit imbalance where the magnitude of policy updates is lower for harder questions. Data-wise, augmentation approaches primarily rephrase questions to enhance diversity without systematically increasing intrinsic difficulty. To address these issues, we propose a two-dual MathForge framework to improve mathematical reasoning by targeting harder questions from both perspectives, which comprises a Difficulty-Aware Group Policy Optimization (DGPO) algorithm and a Multi-Aspect Question Reformulation (MQR) strategy. Specifically, DGPO first rectifies the implicit imbalance in GRPO via difficulty-balanced group advantage estimation, and further prioritizes harder questions by difficulty-aware question-level weighting. Meanwhile, MQR reformulates questions across multiple aspects to increase difficulty while maintaining the original gold answer. Overall, MathForge forms a synergistic loop: MQR expands the data frontier, and DGPO effectively learns from the augmented data. Extensive experiments show that MathForge significantly outperforms existing methods on various mathematical reasoning tasks. The code and augmented data are all available at https://github.com/AMAP-ML/MathForge.
Authors: Boaz Carmeli, Orr Paradise, Shafi Goldwasser, Yonatan Belinkov, Ron Meir
Abstract: We investigate whether \emph{LLM-based agents} can develop task-oriented communication protocols that differ from standard natural language in collaborative reasoning tasks. Our focus is on two core properties such task-oriented protocols may exhibit: Efficiency -- conveying task-relevant information more concisely than natural language, and Covertness -- becoming difficult for external observers to interpret, raising concerns about transparency and control. To investigate these aspects, we use a referential-game framework in which vision-language model (VLM) agents communicate, providing a controlled, measurable setting for evaluating language variants. Experiments show that VLMs can develop effective, task-adapted communication patterns. At the same time, they can develop covert protocols that are difficult for humans and external agents to interpret. We also observe spontaneous coordination between similar models without explicitly shared protocols. These findings highlight both the potential and the risks of task-oriented communication, and position referential games as a valuable testbed for future work in this area.
Authors: Samira Yazdanpourmoghadam, Mahan Balal Pour, Vahid Partovi Nia
Abstract: Combinatorial optimization problems such as the Job-Shop Scheduling Problem (JSP) and Knapsack Problem (KP) are fundamental challenges in operations research, logistics, and eterprise resource planning (ERP). These problems often require sophisticated algorithms to achieve near-optimal solutions within practical time constraints. Recent advances in deep learning have introduced transformer-based architectures as promising alternatives to traditional heuristics and metaheuristics. We leverage the Multi-Type Transformer (MTT) architecture to address these benchmarks in a unified framework. We present an extensive experimental evaluation across standard benchmark datasets for JSP and KP, demonstrating that MTT achieves competitive performance on different size of these benchmark problems. We showcase the potential of multi-type attention on a real application in Ferro-Titanium industry. To the best of our knowledge, we are the first to apply multi-type transformers in real manufacturing.
Authors: Arvid Becker, Pedro Cabalar, Martin Di\'eguez, Susana Hahn, Javier Romero, Torsten Schaub
Abstract: We develop a computational approach to Metric Answer Set Programming (ASP) to allow for expressing quantitative temporal constraints, like durations and deadlines. A central challenge is to maintain scalability when dealing with fine-grained timing constraints, which can significantly exacerbate ASP's grounding bottleneck. To address this issue, we leverage extensions of ASP with difference constraints, a simplified form of linear constraints, to handle time-related aspects externally. Our approach effectively decouples metric ASP from the granularity of time, resulting in a solution that is unaffected by time precision.
Authors: Zishen Wan, Che-Kai Liu, Jiayi Qian, Hanchen Yang, Arijit Raychowdhury, Tushar Krishna
Abstract: Neuro-symbolic AI systems integrate neural perception with symbolic reasoning to enable data-efficient, interpretable, and robust intelligence beyond purely neural models. Although this compositional paradigm has shown superior performance in domains such as reasoning, planning, and verification, its deployment remains challenging due to severe inefficiencies in symbolic and probabilistic inference. Through systematic analysis of representative neuro-symbolic workloads, we identify probabilistic logical reasoning as the inefficiency bottleneck, characterized by irregular control flow, low arithmetic intensity, uncoalesced memory accesses, and poor hardware utilization on CPUs and GPUs. This paper presents REASON, an integrated acceleration framework for probabilistic logical reasoning in neuro-symbolic AI. REASON introduces a unified directed acyclic graph representation that captures common structure across symbolic and probabilistic models, coupled with adaptive pruning and regularization. At the architecture level, REASON features a reconfigurable, tree-based processing fabric optimized for irregular traversal, symbolic deduction, and probabilistic aggregation. At the system level, REASON is tightly integrated with GPU streaming multiprocessors through a programmable interface and multi-level pipeline that efficiently orchestrates compositional execution. Evaluated across six neuro-symbolic workloads, REASON achieves 12-50x speedup and 310-681x energy efficiency over desktop and edge GPUs under TSMC 28 nm node. REASON enables real-time probabilistic logical reasoning, completing end-to-end tasks in 0.8 s with 6 mm2 area and 2.12 W power, demonstrating that targeted acceleration of probabilistic logical reasoning is critical for practical and scalable neuro-symbolic AI and positioning REASON as a foundational system architecture for next-generation cognitive intelligence.
Authors: Vishnu Sashank Dorbala, Dinesh Manocha
Abstract: Foundation models rely on in-context learning for personalized decision making. The limited size of this context window necessitates memory compression and retrieval systems like RAG. These systems however often treat memory as large offline storage spaces, which is unfavorable for embodied agents that are expected to operate under strict memory and compute constraints, online. In this work, we propose MemCtrl, a novel framework that uses Multimodal Large Language Models (MLLMs) for pruning memory online. MemCtrl augments MLLMs with a trainable memory head \mu that acts as a gate to determine which observations or reflections to retain, update, or discard during exploration. We evaluate with training two types of \mu, 1) via an offline expert, and 2) via online RL, and observe significant improvement in overall embodied task completion ability on \mu-augmented MLLMs. In particular, on augmenting two low performing MLLMs with MemCtrl on multiple subsets of the EmbodiedBench benchmark, we observe that \mu-augmented MLLMs show an improvement of around 16% on average, with over 20% on specific instruction subsets. Finally, we present a qualitative analysis on the memory fragments collected by \mu, noting the superior performance of \mu augmented MLLMs on long and complex instruction types.
Authors: Saurav Prateek
Abstract: This paper introduces a novel Deep Researcher architecture designed to generate detailed research reports on complex PhD level topics by addressing the inherent limitations of the Parallel Scaling paradigm. Our system utilizes two key innovations: Sequential Research Plan Refinement via Reflection and a Candidates Crossover algorithm. The sequential refinement process is demonstrated as an efficient method that allows the agent to maintain a centralized Global Research Context, enabling it to look back at current progress, reason about the research plan, and intelligently make changes at runtime. This dynamic adaptation contrasts with parallel approaches, which often suffer from siloed knowledge. The Candidates Crossover algorithm further enhances search efficiency by deploying multiple LLM candidates with varied parameters to explore a larger search space, with their findings synthesized to curate a comprehensive final research response. The process concludes with One Shot Report Generation, ensuring the final document is informed by a unified narrative and high fact density. Powered by the Gemini 2.5 Pro model, our Deep Researcher was evaluated on the DeepResearch Bench, a globally recognized benchmark of 100 doctoral level research tasks. Our architecture achieved an overall score of 46.21, demonstrating superior performance by surpassing leading deep research agents such as Claude Researcher, Nvidia AIQ Research Assistant, Perplexity Research, Kimi Researcher and Grok Deeper Search present on the DeepResearch Bench actively running leaderboard. This performance marginally exceeds our previous work, Static DRA, and reinforces the finding that sequential scaling consistently outperforms the parallel self consistency paradigm.
Authors: Sebastiano Monti, Carlo Nicolini, Gianni Pellegrini, Jacopo Staiano, Bruno Lepri
Abstract: Although the capabilities of large language models have been increasingly tested on complex reasoning tasks, their long-horizon planning abilities have not yet been extensively investigated. In this work, we provide a systematic assessment of the planning and long-horizon reasoning capabilities of state-of-the-art Large Reasoning Models (LRMs). We propose a novel benchmark based on Sokoban puzzles, intentionally simplified to isolate long-horizon planning from state persistence. Our findings reveal a consistent degradation in planning performance when more than 25 moves are required to reach the solution, suggesting a fundamental constraint on forward planning capacity. We show that equipping LRMs with Planning Domain Definition Language (PDDL) parsing, validation, and solving tools allows for modest improvements, suggesting inherent architectural limitations which might not be overcome by test-time scaling approaches alone.
Authors: Saeid Rajabi, Chengmo Yang, Satwik Patnaik
Abstract: Formal Verification (FV) relies on high-quality SystemVerilog Assertions (SVAs), but the manual writing process is slow and error-prone. Existing LLM-based approaches either generate assertions from scratch or ignore structural patterns in hardware designs and expert-crafted assertions. This paper presents STELLAR, the first framework that guides LLM-based SVA generation with structural similarity. STELLAR represents RTL blocks as AST structural fingerprints, retrieves structurally relevant (RTL, SVA) pairs from a knowledge base, and integrates them into structure-guided prompts. Experiments show that STELLAR achieves superior syntax correctness, stylistic alignment, and functional correctness, highlighting structure-aware retrieval as a promising direction for industrial FV.
Authors: Ziyu Hu, Zhiqing Zhong, Weijian Zheng, Zhijing Ye, Xuwei Tan, Xueru Zhang, Zheng Xie, Rajkumar Kettimuthu, Xiaodong Yu
Abstract: The exponential growth of large language models has outpaced the capabilities of traditional CPU and GPU architectures due to the slowdown of Moore's Law. Dataflow AI accelerators present a promising alternative; however, there remains a lack of in-depth performance analysis and standardized benchmarking methodologies for LLM training. We introduce DABench-LLM, the first benchmarking framework designed for evaluating LLM workloads on dataflow-based accelerators. By combining intra-chip performance profiling and inter-chip scalability analysis, DABench-LLM enables comprehensive evaluation across key metrics such as resource allocation, load balance, and resource efficiency. The framework helps researchers rapidly gain insights into underlying hardware and system behaviors, and provides guidance for performance optimizations. We validate DABench-LLM on three commodity dataflow accelerators, Cerebras WSE-2, SambaNova RDU, and Graphcore IPU. Our framework reveals performance bottlenecks and provides specific optimization strategies, demonstrating its generality and effectiveness across a diverse range of dataflow-based AI hardware platforms.
Authors: Jingxin Wang, Shitong Guo, Ruicheng Dai, Wenhui Liang, Ruogu Ding, Xin Ning, Weikang Qian
Abstract: Targeting error-tolerant applications, approximate circuits introduce controlled errors to significantly improve performance, power, and area (PPA) of circuits. In this work, we introduce GTAC, a novel generative Transformer-based model for producing approximate circuits. By leveraging principles of approximate computing and AI-driven EDA, our model innovatively integrates error thresholds into the design process. Experimental results show that compared with a state-of-the-art method, GTAC further reduces 6.4% area under the error rate constraint, while being 4.3x faster.
Authors: Duo Chai, Zizhen Liu, Shuhuai Wang, Songwei Pei, Cheng Liu, Huawei Li, Shangguang Wang
Abstract: Large language models (LLMs) are highly compute- and memory-intensive, posing significant demands on high-performance GPUs. At the same time, advances in GPU technology driven by shrinking transistor sizes and lower operating voltages have made these devices increasingly susceptible to soft errors. While prior work has examined GPU reliability, most studies have focused on general-purpose applications or conventional neural networks mostly used for vision tasks such as classification and detection. In contrast, systematic analysis of modern large-scale LLMs remains limited, despite their rapid adoption in diverse application scenarios. Given the unique characteristics of LLMs, their resilience to soft errors may differ substantially from earlier models. To bridge this gap, we conduct the first instruction-level fault injection study of LLM inference. Our approach reveals reliability characteristics from multiple perspectives, highlighting the effects of model architecture, parameter scale, and task complexity. These findings provide new insights into LLM reliability and inform the design of more effective fault tolerance mechanisms.
Authors: Shinwoo Park, Yo-Sub Han
Abstract: Distinguishing human-written Korean text from fluent LLM outputs remains difficult even for linguistically trained readers, who can over-trust surface well-formedness. We study whether expert detection can be treated as a learnable skill and improved through structured calibration. We introduce LREAD, a rubric derived from national Korean writing standards and adapted to target micro-level artifacts (e.g., punctuation optionality, spacing behavior, and register shifts). In a three-phase longitudinal blind protocol with Korean linguistics majors, Phase 1 measures intuition-only detection, Phase 2 enforces criterion-level scoring with explicit justifications, and Phase 3 evaluates domain-focused mastery on held-out elementary essays. Across phases, majority-vote accuracy increases from 60% to 100%, accompanied by stronger inter-annotator agreement (Fleiss' kappa: -0.09 --> 0.82). Compared to state-of-the-art LLM detectors, calibrated humans rely more on language-specific micro-diagnostics that are not well captured by coarse discourse priors. Our findings suggest that rubric-scaffolded expert judgment can serve as an interpretable complement to automated detectors for non-English settings, and we release the full rubric and a taxonomy of calibrated detection signatures.
Authors: Maxwell Crouse, Ibrahim Abdelaziz, Kshitij Fadnis, Siva Sankalp Patel, Kinjal Basu, Chulaka Gunasekara, Sadhana Kumaravel, Asim Munawar, Pavan Kapanipathi
Abstract: Synthetic data has proven itself to be a valuable resource for tuning smaller, cost-effective language models to handle the complexities of multi-turn tool calling conversations. While many frameworks and systems for producing synthetic multi-turn tool calling data have been proposed, prior works have frequently assumed that any tool calling interactions will take place in an execution environment that maintains state. When such an environment is available, this is advantageous as it allows for the validity of an interaction to be determined by whether or not the state of the execution environment matches to some prespecified objective. Unfortunately, this does not hold in many real-world tool use settings, e.g., in enterprise settings where data security is of the utmost importance or in cases where tool specifications are synthesized from multiple sources. In this work, we address this gap by introducing a data generation method, DiGiT-TC, that is designed to produce tool calling conversations that have the characteristics of conversations generated through search in a stateful environment. The key to our technique lies in a novel generation pattern that allows our approach to implicitly represent certain tool calls in the user request. We validate our approach on standard tool calling benchmarks and demonstrate that, even in stateful problem settings, our approach results in strong performance gains.
Authors: Paul Tarau
Abstract: We introduce the \emph{Arrow Language Model}, a neural architecture derived from an intuitionistic-logic interpretation of next-token prediction. Instead of representing tokens as additive embeddings mixed by attention, we encode a prefix as a \emph{left-nested implication chain} whose structure preserves order through non-commutative composition. Next-token prediction corresponds to \emph{modus ponens}, and sequence processing becomes constructive proof extension under the Curry--Howard correspondence. Our Prolog-based specialized theorem provers validate fundamental properties of the neural models, among which relations between commutative vs. non-commutative sequencing and single-token vs. multi-token prediction choices. We show that a neural architecture equivalent to multiplicative RNNs arises naturally from a proof-theoretic interpretation of next-token prediction as nested intuitionistic implication, we present a practical low-rank neural realization and position the model relative to Transformers and state-space models. Keywords: logic-based derivation of neural architectures, intuitionistic implicational logic, token-as-operator neural models, state-space models, alternatives to transformer-based foundational models.
Authors: Junseok Lee, Nahoon Kim, Sangyong Lee, Chang-Jae Chun
Abstract: Knowledge distillation is one of the most effective methods for model compression. Previous studies have focused on the student model effectively training the predictive distribution of the teacher model. However, during training, the student model may inherit the shortcomings of the teacher model, which can lead to a decline in generalization capacity. To mitigate this issue, we propose adaptive self-knowledge distillation (ASKD), which dynamically reduces the dependence of the teacher model to improve the self-training capacity, and performs the self-knowledge distillation method to improve the generalization capacity of the student model. We further distill the Whisper model into a smaller variant, called FastWhisper. In our post-training setting, FastWhisper achieved a word error rate of 1.07% lower than the teacher model Whisper, and its relative inference time was 5 times faster.
Authors: Xiaochen Zhu, Caiqi Zhang, Yizhou Chi, Tom Stafford, Nigel Collier, Andreas Vlachos
Abstract: Multi-agent debate (MAD) is widely used to improve large language model (LLM) performance through test-time scaling, yet recent work shows that vanilla MAD often underperforms simple majority vote despite higher computational cost. Studies show that, under homogeneous agents and uniform belief updates, debate preserves expected correctness and therefore cannot reliably improve outcomes. Drawing on findings from human deliberation and collective decision-making, we identify two key mechanisms missing from vanilla MAD: (i) diversity of initial viewpoints and (ii) explicit, calibrated confidence communication. We propose two lightweight interventions. First, a diversity-aware initialisation that selects a more diverse pool of candidate answers, increasing the likelihood that a correct hypothesis is present at the start of debate. Second, a confidence-modulated debate protocol in which agents express calibrated confidence and condition their updates on others' confidence. We show theoretically that diversity-aware initialisation improves the prior probability of MAD success without changing the underlying update dynamics, while confidence-modulated updates enable debate to systematically drift to the correct hypothesis. Empirically, across six reasoning-oriented QA benchmarks, our methods consistently outperform vanilla MAD and majority vote. Our results connect human deliberation with LLM-based debate and demonstrate that simple, principled modifications can substantially enhance debate effectiveness.
Authors: Laya Iyer, Kriti Aggarwal, Sanmi Koyejo, Gail Heyman, Desmond C. Ong, Subhabrata Mukherjee
Abstract: Supportive conversation depends on skills that go beyond language fluency, including reading emotions, adjusting tone, and navigating moments of resistance, frustration, or distress. Despite rapid progress in language models, we still lack a clear way to understand how their abilities in these interpersonal domains compare to those of humans. We introduce HEART, the first-ever framework that directly compares humans and LLMs on the same multi-turn emotional-support conversations. For each dialogue history, we pair human and model responses and evaluate them through blinded human raters and an ensemble of LLM-as-judge evaluators. All assessments follow a rubric grounded in interpersonal communication science across five dimensions: Human Alignment, Empathic Responsiveness, Attunement, Resonance, and Task-Following. HEART uncovers striking behavioral patterns. Several frontier models approach or surpass the average human responses in perceived empathy and consistency. At the same time, humans maintain advantages in adaptive reframing, tension-naming, and nuanced tone shifts, particularly in adversarial turns. Human and LLM-as-judge preferences align on about 80 percent of pairwise comparisons, matching inter-human agreement, and their written rationales emphasize similar HEART dimensions. This pattern suggests an emerging convergence in the criteria used to assess supportive quality. By placing humans and models on equal footing, HEART reframes supportive dialogue as a distinct capability axis, separable from general reasoning or linguistic fluency. It provides a unified empirical foundation for understanding where model-generated support aligns with human social judgment, where it diverges, and how affective conversational competence scales with model size.
Authors: Boxiang Zhao, Qince Li, Zhonghao Wang, Zelin Cao, Yi Wang, Peng Cheng, Bo Lin
Abstract: As Large Language Models (LLMs) evolve into autonomous agents, the capability to faithfully translate natural language into rigorous structured formats-essential for tool invocation-and to convert complex tabular information into machine-readable specifications has become paramount. However, current evaluations lack effective methodologies to measure this structural fidelity without costly human intervention, as traditional text metrics fail to detect semantic drift in code-like outputs. This paper proposes Table-BiEval, a novel approach based on a human-free, self-supervised evaluation framework, to assess LLMs performance quantitatively. By leveraging deterministic Intermediate Representations, our framework calculates Content Semantic Accuracy and Normalized Tree Edit Distance to decouple structure from content. Also, it empirically evaluates 15 state-of-the-art LLMs across dual topological dimensions-hierarchical structures and flat tables. The results reveal substantial variability, highlighting that mid-sized models can surprisingly outperform larger counterparts in structural efficiency and confirming that deep recursive nesting remains a universal bottleneck for current architectures.
Authors: Yitian Chen, Cheng Cheng, Yinan Sun, Zi Ling, Dongdong Ge
Abstract: Large Language Models (LLMs) have demonstrated impressive progress in optimization modeling, fostering a rapid expansion of new methodologies and evaluation benchmarks. However, the boundaries of their capabilities in automated formulation and problem solving remain poorly understood, particularly when extending to complex, real-world tasks. To bridge this gap, we propose OPT-ENGINE, an extensible benchmark framework designed to evaluate LLMs on optimization modeling with controllable and scalable difficulty levels. OPT-ENGINE spans 10 canonical tasks across operations research, with five Linear Programming and five Mixed-Integer Programming. Utilizing OPT-ENGINE, we conduct an extensive study of LLMs' reasoning capabilities, addressing two critical questions: 1.) Do LLMs' performance remain robust when generalizing to out-of-distribution optimization tasks that scale in complexity beyond current benchmark levels? and 2.) At what stage, from problem interpretation to solution generation, do current LLMs encounter the most significant bottlenecks? Our empirical results yield two key insights: first, tool-integrated reasoning with external solvers exhibits significantly higher robustness as task complexity escalates, while pure-text reasoning reaches a ceiling; second, the automated formulation of constraints constitutes the primary performance bottleneck. These findings provide actionable guidance for developing next-generation LLMs for advanced optimization. Our code is publicly available at \textcolor{blue}{https://github.com/Cardinal-Operations/OPTEngine}.
Authors: Yinuo Liu, Emre Sezgin, Eric A. Youngstrom
Abstract: Introduction: Large language models (LLMs) can process requests and generate texts, but their feasibility for assessing complex academic content needs further investigation. To explore LLM's potential in assisting scientific review, this study examined ChatGPT-5, Gemini-3-Pro, and Claude-Sonnet-4.5's consistency and reliability in evaluating abstracts compared to one another and to human reviewers. Methods: 160 abstracts from a local conference were graded by human reviewers and three LLMs using one rubric. Composite score distributions across three LLMs and fourteen reviewers were examined. Inter-rater reliability was calculated using intraclass correlation coefficients (ICCs) for within-AI reliability and AI-human concordance. Bland-Altman plots were examined for visual agreement patterns and systematic bias. Results: LLMs achieved good-to-excellent agreement with each other (ICCs: 0.59-0.87). ChatGPT and Claude reached moderate agreement with human reviewers on overall quality and content-specific criteria, with ICCs ~.45-.60 for composite, impression, clarity, objective, and results. They exhibited fair agreement on subjective dimensions, with ICC ranging from 0.23-0.38 for impact, engagement, and applicability. Gemini showed fair agreement on half criteria and no reliability on impact and applicability. Three LLMs showed acceptable or negligible mean difference (ChatGPT=0.24, Gemini=0.42, Claude=-0.02) from the human mean composite scores. Discussion: LLMs could process abstracts in batches with moderate agreement with human experts on overall quality and objective criteria. With appropriate process architecture, they can apply a rubric consistently across volumes of abstracts exceeding feasibility for a human rater. The weaker performance on subjective dimensions indicates that AI should serve a complementary role in evaluation, while human expertise remains essential.
Authors: Nora Graichen, Iria de-Dios-Flores, Gemma Boleda
Abstract: We present a systematic review of 337 articles evaluating the syntactic abilities of Transformer-based language models, reporting on 1,015 model results from a range of syntactic phenomena and interpretability methods. Our analysis shows that the state of the art presents a healthy variety of methods and data, but an over-focus on a single language (English), a single model (BERT), and phenomena that are easy to get at (like part of speech and agreement). Results also suggest that TLMs capture these form-oriented phenomena well, but show more variable and weaker performance on phenomena at the syntax-semantics interface, like binding or filler-gap dependencies. We provide recommendations for future work, in particular reporting complete data, better aligning theoretical constructs and methods across studies, increasing the use of mechanistic methods, and broadening the empirical scope regarding languages and linguistic phenomena.
Authors: David Linus Ostby
Abstract: We introduce Stingy Context, a hierarchical tree-based compression scheme achieving 18:1 reduction in LLM context for auto-coding tasks. Using our TREEFRAG exploit decomposition, we reduce a real source code base of 239k tokens to 11k tokens while preserving task fidelity. Empirical results across 12 Frontier models show 94 to 97% success on 40 real-world issues at low cost, outperforming flat methods and mitigating lost-in-the-middle effects.
Authors: Jacob Nielsen, Stine L. Beltoft, Peter Schneider-Kamp, Lukas Galke Poech
Abstract: We introduce a new benchmark for Danish culture via cultural heritage, Daisy, based on the curated topics from the Danish Culture Canon 2006. For each artifact in the culture canon, we query the corresponding Wikipedia page and have a language model generate random questions. This yields a sampling strategy within each work, with a mix of central of peripheral questions for each work, not only knowledge of mainstream information, but also in-depth cornerstones defining the heritage of Danish Culture, defined by the Canon committee. Each question-answer pair is humanly approved or corrected in the final dataset consisting of 741 close-ended question answer pairs covering topics, from 1300 BC. archaeological findings, 1700 century poems and musicals pieces to contemporary pop music and Danish design and architecture.
Authors: Kei Saito
Abstract: Non-Resolution Reasoning (NRR) provides a formal framework for maintaining semantic ambiguity rather than forcing premature interpretation collapse. While the foundational architecture establishes state spaces and operators for ambiguity-preserving computation, the critical question of how natural language maps to these mathematical structures remains open. This paper introduces the text-to-state mapping function {\phi} that transforms linguistic input into superposition states within the NRR framework. We formalize the Contradiction-Preservation Principle, which requires that genuinely ambiguous expressions maintain non-zero entropy in their state representations, and develop extraction protocols using existing Large Language Models as interpretation generators. Empirical validation across 68 test sentences spanning lexical, structural, and pragmatic ambiguity demonstrates that our mapping achieves mean Shannon entropy H(S) = 1.087 bits for ambiguous inputs while baseline single-interpretation approaches yield H(S) = 0.000. The framework provides the missing algorithmic bridge between raw text and the formal state spaces on which NRR operators act, enabling architectural collapse deferment in language model inference.
Authors: Claire Nicholson
Abstract: Large language models (LLMs) are widely used for tasks ranging from summarisation to decision support. In practice, identical prompts do not always produce identical outputs, even when temperature and other decoding parameters are fixed. In this work, we conduct repeated-run experiments to empirically quantify baseline behavioural drift, defined as output variability observed when the same prompt is issued multiple times under operator-free conditions. We evaluate two publicly accessible models, gpt-4o-mini and llama3.1-8b, across five prompt categories using exact repeats, perturbed inputs, and reuse modes at temperatures of 0.0 and 0.7. Drift is measured using unique output fractions, lexical similarity, and word count statistics, enabling direct comparison across models, prompting modes, and deployment types. The results show that nondeterminism persists even at temperature 0.0, with distinct variability patterns by model size, deployment, and prompt type. We situate these findings within existing work on concept drift, behavioural drift, and infrastructure-induced nondeterminism, discuss the limitations of lexical metrics, and highlight emerging semantic approaches. By establishing a systematic empirical baseline in the absence of stabilisation techniques, this study provides a reference point for evaluating future drift mitigation and control methods.
Authors: Yiting Shen, Kun Li, Wei Zhou, Songlin Hu
Abstract: Large Language Model (LLM)-based agents are increasingly deployed for complex, tool-based tasks where long-term memory is critical to driving actions. Existing benchmarks, however, primarily test a angent's ability to passively retrieve isolated facts in response to explicit questions. They fail to evaluate the more crucial capability of actively applying memory to execute tasks. To address this gap, we introduce \textsc{Mem2ActBench}, a benchmark for evaluating whether agents can proactively leverage long-term memory to execute tool-based actions by selecting appropriate tools and grounding their parameters. The benchmark simulates persistent assistant usage, where users mention the same topic across long, interrupted interactions and expect previously established preferences and task states to be implicitly applied. We build the dataset with an automated pipeline that merges heterogeneous sources (ToolACE, BFCL, Oasst1), resolves conflicts via consistency modeling, and synthesizes 2,029 sessions with 12 user--assistant--tool turns on average. From these memory chains, a reverse-generation method produces 400 tool-use tasks, with human evaluation confirming 91.3\% are strongly memory-dependent. Experiments on seven memory frameworks show that current systems remain inadequate at actively utilizing memory for parameter grounding, highlighting the need for more effective approaches to evaluate and improve memory application in task execution.
Authors: Minseo Kwak, Jaehyung Kim
Abstract: The opacity of massive pretraining corpora in Large Language Models (LLMs) raises significant privacy and copyright concerns, making pretraining data detection a critical challenge. Existing state-of-the-art methods typically rely on token likelihoods, yet they often overlook the divergence from the model's top-1 prediction and local correlation between adjacent tokens. In this work, we propose Gap-K%, a novel pretraining data detection method grounded in the optimization dynamics of LLM pretraining. By analyzing the next-token prediction objective, we observe that discrepancies between the model's top-1 prediction and the target token induce strong gradient signals, which are explicitly penalized during training. Motivated by this, Gap-K% leverages the log probability gap between the top-1 predicted token and the target token, incorporating a sliding window strategy to capture local correlations and mitigate token-level fluctuations. Extensive experiments on the WikiMIA and MIMIR benchmarks demonstrate that Gap-K% achieves state-of-the-art performance, consistently outperforming prior baselines across various model sizes and input lengths.
Authors: Adnan Ahmad, Chiara Boldrini, Lorenzo Valerio, Andrea Passarella, Marco Conti
Abstract: Decentralized Federated Learning (DFL) is a serverless collaborative machine learning paradigm where devices collaborate directly with neighbouring devices to exchange model information for learning a generalized model. However, variations in individual experiences and different levels of device interactions lead to data and model initialization heterogeneities across devices. Such heterogeneities leave variations in local model parameters across devices that leads to slower convergence. This paper tackles the data and model heterogeneity by explicitly addressing the parameter level varying evidential credence across local models. A novel aggregation approach is introduced that captures these parameter variations in local models and performs robust aggregation of neighbourhood local updates. Specifically, consensus weights are generated via approximation of second-order information of local models on their local datasets. These weights are utilized to scale neighbourhood updates before aggregating them into global neighbourhood representation. In extensive experiments with computer vision tasks, the proposed approach shows strong generalizability of local models at reduced communication costs.
Authors: Tobias Habermann, Michael Mecik, Zhenyu Wang, C\'esar David Vera, Martin Kumm, Mario Garrido
Abstract: Among hardware accelerators for deep-learning inference, data flow implementations offer low latency and high throughput capabilities. In these architectures, each neuron is mapped to a dedicated hardware unit, making them well-suited for field-programmable gate array (FPGA) implementation. Previous unrolled implementations mostly focus on fully connected networks because of their simplicity, although it is well known that convolutional neural networks (CNNs) require fewer computations for the same accuracy. When observing the data flow in CNNs, pooling layers and convolutional layers with a stride larger than one, the number of data at their output is reduced with respect to their input. This data reduction strongly affects the data rate in a fully parallel implementation, making hardware units heavily underutilized unless it is handled properly. This work addresses this issue by analyzing the data flow of CNNs and presents a novel approach to designing data-rate-aware, continuous-flow CNN architectures. The proposed approach ensures a high hardware utilization close to 100% by interleaving low data rate signals and sharing hardware units, as well as using the right parallelization to achieve the throughput of a fully parallel implementation. The results show that a significant amount of the arithmetic logic can be saved, which allows implementing complex CNNs like MobileNet on a single FPGA with high throughput.
Authors: M Zafir Sadik Khan, Kimia Azar, Hadi Kamali
Abstract: In last two years, large language models (LLMs) have shown strong capabilities in code generation, including hardware design at register-transfer level (RTL). While their use in high-level synthesis (HLS) remains comparatively less mature, the ratio of HLS- to RTL-focused studies has shifted from 1:10 to 2:10 in the past six months, indicating growing interest in leveraging LLMs for high-level design entry while relying on downstream synthesis for optimization. This growing trend highlights the need for a comprehensive benchmarking and evaluation framework dedicated to LLM-based HLS. To address this, We present Bench4HLS for evaluating LLM-generated HLS designs. Bench4HLS comprises 170 manually drafted and validated case studies, spanning small kernels to complex accelerators, curated from widely used public repositories. The framework supports fully automated assessment of compilation success, functional correctness via simulation, and synthesis feasibility/optimization. Crucially, Bench4HLS integrates a pluggable API for power, performance, and area (PPA) analysis across various HLS toolchains and architectures, demonstrated here with Xilinx Vitis HLS and validated on Catapult HLS. By providing a structured, extensible, and plug-and-play testbed, Bench4HLS establishes a foundational methodology for benchmarking LLMs in HLS workflows.
Authors: Thomas Schuster, Julius Tr\"ogele, Nico D\"oring, Robin Kr\"uger, Matthieu Hoffmann, Holger Friedrich
Abstract: Automatic Speech Recognition (ASR) offers significant potential to reduce the workload of medical personnel, for example, through the automation of documentation tasks. While numerous benchmarks exist for the English language, specific evaluations for the German-speaking medical context are still lacking, particularly regarding the inclusion of dialects. In this article, we present a curated dataset of simulated doctor-patient conversations and evaluate a total of 29 different ASR models. The test field encompasses both open-weights models from the Whisper, Voxtral, and Wav2Vec2 families as well as commercial state-of-the-art APIs (AssemblyAI, Deepgram). For evaluation, we utilize three different metrics (WER, CER, BLEU) and provide an outlook on qualitative semantic analysis. The results demonstrate significant performance differences between the models: while the best systems already achieve very good Word Error Rates (WER) of partly below 3%, the error rates of other models, especially concerning medical terminology or dialect-influenced variations, are considerably higher.
Authors: Jiayu Xu, Junbiao Pang
Abstract: Learning from Noisy Labels (LNL) presents a fundamental challenge in deep learning, as real-world datasets often contain erroneous or corrupted annotations, \textit{e.g.}, data crawled from Web. Current research focuses on sophisticated label correction mechanisms. In contrast, this paper adopts a novel perspective by establishing a theoretical analysis the relationship between flatness of the loss landscape and the presence of label noise. In this paper, we theoretically demonstrate that carefully simulated label noise synergistically enhances both the generalization performance and robustness of label noises. Consequently, we propose Noise-Compensated Sharpness-aware Minimization (NCSAM) to leverage the perturbation of Sharpness-Aware Minimization (SAM) to remedy the damage of label noises. Our analysis reveals that the testing accuracy exhibits a similar behavior that has been observed on the noise-clear dataset. Extensive experimental results on multiple benchmark datasets demonstrate the consistent superiority of the proposed method over existing state-of-the-art approaches on diverse tasks.
Authors: Wenhao Zou, Yuwei Miao, Zhanyu Ma, Jun Xu, Jiuchong Gao, Jinghua Hao, Renqing He, Jingwen Xu
Abstract: Real-time voice agents face a dilemma: end-to-end models often lack deep reasoning, while cascaded pipelines incur high latency by executing ASR, LLM reasoning, and TTS strictly in sequence, unlike human conversation where listeners often start thinking before the speaker finishes. Since cascaded architectures remain the dominant choice for complex tasks, existing cascaded streaming strategies attempt to reduce this latency via mechanical segmentation (e.g., fixed chunks, VAD-based splitting) or speculative generation, but they frequently either break semantic units or waste computation on predictions that must be rolled back. To address these challenges, we propose LTS-VoiceAgent, a Listen-Think-Speak framework that explicitly separates when to think from how to reason incrementally. It features a Dynamic Semantic Trigger to detect meaningful prefixes, and a Dual-Role Stream Orchestrator that coordinates a background Thinker (for state maintenance) and a foreground Speaker (for speculative solving). This parallel design enables "thinking while speaking" without blocking responses. We also introduce a Pause-and-Repair benchmark containing natural disfluencies to stress-test streaming robustness. Experiments across VERA, Spoken-MQA, BigBenchAudio, and our benchmark show that LTS-VoiceAgent achieves a stronger accuracy-latency-efficiency trade-off than serial cascaded baselines and existing streaming strategies.
Authors: Ibrahim Albulushi, Saleh Bunaiyan, Suraj S. Cheema, Hesham ElSawy, Feras Al-Dirini
Abstract: Extending the intelligence of sensors to the data-acquisition process - deciding whether to sample or not - can result in transformative energy-efficiency gains. However, making such a decision in a deterministic manner involves risk of losing information. Here we present a sensing paradigm that enables making such a decision in a probabilistic manner. The paradigm takes inspiration from the autonomous nervous system and employs a probabilistic neuron (p-neuron) driven by an analog feature extraction circuit. The response time of the system is on the order of microseconds, over-coming the sub-sampling-rate response time limit and enabling real-time intelligent autonomous activation of data-sampling. Validation experiments on active seismic survey data demonstrate lossless probabilistic data acquisition, with a normalized mean squared error of 0.41%, and 93% saving in the active operation time of the system and the number of generated samples.
Authors: Yuxiang Wang, Hongyu Liu, Dekun Chen, Xueyao Zhang, Zhizheng Wu
Abstract: As Speech Language Models (SLMs) transition from personal devices to shared, multi-user environments such as smart homes, a new challenge emerges: the model is expected to distinguish between users to manage information flow appropriately. Without this capability, an SLM could reveal one user's confidential schedule to another, a privacy failure we term interactional privacy. Thus, the ability to generate speaker-aware responses becomes essential for SLM safe deployment. Current SLM benchmarks test dialogue ability but overlook speaker identity. Multi-speaker benchmarks check who said what without assessing whether SLMs adapt their responses. Privacy benchmarks focus on globally sensitive data (e.g., bank passwords) while neglecting contextual privacy-sensitive information (e.g., a user's private appointment). To address this gap, we introduce VoxPrivacy, the first benchmark designed to evaluate interactional privacy in SLMs. VoxPrivacy spans three tiers of increasing difficulty, from following direct secrecy commands to proactively protecting privacy. Our evaluation of nine SLMs on a 32-hour bilingual dataset reveals a widespread vulnerability: most open-source models perform close to random chance (around 50% accuracy) on conditional privacy decisions, while even strong closed-source systems fall short on proactive privacy inference. We further validate these findings on Real-VoxPrivacy, a human-recorded subset, confirming that failures observed on synthetic data persist in real speech. Finally, we demonstrate a viable path forward: by fine-tuning on a new 4,000-hour training set, we improve privacy-preserving abilities while maintaining robustness. To support future work, we release the VoxPrivacy benchmark, the large-scale training set, and the fine-tuned model to foster the development of safer and more context-aware SLMs.
Authors: Youness Dkhissi (LIUM), Valentin Vielzeuf (LIUM), Elys Allesiardo (LIUM), Anthony Larcher (LIUM)
Abstract: Transformer-based architectures are the most used architectures in many deep learning fields like Natural Language Processing, Computer Vision or Speech processing. It may encourage the direct use of Transformers in the constrained tasks, without questioning whether it will yield the same benefits as in standard tasks. Given specific constraints, it is essential to evaluate the relevance of transformer models. This work questions the suitability of transformers for specific domains. We argue that the high computational requirements and latency issues associated with these models do not align well with streaming applications. Our study promotes the search for alternative strategies to improve efficiency without sacrificing performance. In light of this observation, our paper critically examines the usefulness of transformer architecture in such constrained environments. As a first attempt, we show that the computational cost for Streaming Automatic Speech Recognition (ASR) can be reduced using deformable convolution instead of Self-Attention. Furthermore, we show that Self-Attention mechanisms can be entirely removed and not replaced, without observing significant degradation in the Word Error Rate.
Authors: Huanlin Gao, Ping Chen, Fuyuan Shi, Ruijia Wu, Li YanTao, Qiang Hui, Yuren You, Ting Lu, Chao Tan, Shaoan Zhao, Zhaoxiang Liu, Fang Zhao, Kai Wang, Shiguo Lian
Abstract: We present MeanCache, a training-free caching framework for efficient Flow Matching inference. Existing caching methods reduce redundant computation but typically rely on instantaneous velocity information (e.g., feature caching), which often leads to severe trajectory deviations and error accumulation under high acceleration ratios. MeanCache introduces an average-velocity perspective: by leveraging cached Jacobian--vector products (JVP) to construct interval average velocities from instantaneous velocities, it effectively mitigates local error accumulation. To further improve cache timing and JVP reuse stability, we develop a trajectory-stability scheduling strategy as a practical tool, employing a Peak-Suppressed Shortest Path under budget constraints to determine the schedule. Experiments on FLUX.1, Qwen-Image, and HunyuanVideo demonstrate that MeanCache achieves 4.12X and 4.56X and 3.59X acceleration, respectively, while consistently outperforming state-of-the-art caching baselines in generation quality. We believe this simple yet effective approach provides a new perspective for Flow Matching inference and will inspire further exploration of stability-driven acceleration in commercial-scale generative models.
Authors: Canyang Zhao, Bolin Peng, J. Patrick Mayo, Ce Ju, Bing Liu
Abstract: Cross-session nonstationarity in neural activity recorded by implanted electrodes is a major challenge for invasive Brain-computer interfaces (BCIs), as decoders trained on data from one session often fail to generalize to subsequent sessions. This issue is further exacerbated in practice, as retraining or adapting decoders becomes particularly challenging when only limited data are available from a new session. To address this challenge, we propose a Task-Conditioned Latent Alignment framework (TCLA) for cross-session neural decoding. Building upon an autoencoder architecture, TCLA first learns a low-dimensional representation of neural dynamics from a source session with sufficient data. For target sessions with limited data, TCLA then aligns target latent representations to the source in a task-conditioned manner, enabling effective transfer of learned neural dynamics. We evaluate TCLA on the macaque motor and oculomotor center-out dataset. Compared to baseline methods trained solely on target-session data, TCLA consistently improves decoding performance across datasets and decoding settings, with gains in the coefficient of determination of up to 0.386 for y coordinate velocity decoding in a motor dataset. These results suggest that TCLA provides an effective strategy for transferring knowledge from source to target sessions, enabling more robust neural decoding under conditions with limited data.
Authors: Jinlin Liu, Wei Chen, Xiaojin Zhang
Abstract: Collecting web data to train deep models has become increasingly common, raising concerns about unauthorized data usage. To mitigate this issue, unlearnable examples introduce imperceptible perturbations into data, preventing models from learning effectively. However, existing methods typically rely on deep neural networks as surrogate models for perturbation generation, resulting in significant computational costs. In this work, we propose Perturbation-Induced Linearization (PIL), a computationally efficient yet effective method that generates perturbations using only linear surrogate models. PIL achieves comparable or better performance than existing surrogate-based methods while reducing computational time dramatically. We further reveal a key mechanism underlying unlearnable examples: inducing linearization to deep models, which explains why PIL can achieve competitive results in a very short time. Beyond this, we provide an analysis about the property of unlearnable examples under percentage-based partial perturbation. Our work not only provides a practical approach for data protection but also offers insights into what makes unlearnable examples effective.
Authors: Micha{\l} Gromadzki, Anna Wr\'oblewska, Agnieszka Kaliska
Abstract: The rapid progress of large language models has enabled the generation of text that closely resembles human writing, creating challenges for authenticity verification in education, publishing, and digital security. Detecting AI-generated text has therefore become a crucial technical and ethical issue. This paper presents a comprehensive study of AI-generated text detection based on large-scale corpora and novel training strategies. We introduce a 1-billion-token corpus of human-authored texts spanning multiple genres and a 1.9-billion-token corpus of AI-generated texts produced by prompting a variety of LLMs across diverse domains. Using these resources, we develop and evaluate numerous detection models and propose two novel training paradigms: Per LLM and Per LLM family fine-tuning. Across a 100-million-token benchmark covering 21 large language models, our best fine-tuned detector achieves up to $99.6\%$ token-level accuracy, substantially outperforming existing open-source baselines.
Authors: J. Ben Tamo, Daniel Carlander-Reuterfelt, Jonathan Rubin, Dezhi Hong, Mingxian Wang, Oleg Poliannikov
Abstract: Despite multilingual pretraining, large language models often struggle with non-English tasks, particularly in language control, the ability to respond in the intended language. We identify and characterize two key failure modes: the multilingual transfer bottleneck (correct language, incorrect task response) and the language consistency bottleneck (correct task response, wrong language). To systematically surface these issues, we design a four-scenario evaluation protocol spanning MMLU, MGSM, and XQuAD benchmarks. To probe these issues with interpretability, we extend logit lens analysis to track language probabilities layer by layer and compute cross-lingual semantic similarity of hidden states. The results reveal a three-phase internal structure: early layers align inputs into a shared semantic space, middle layers perform task reasoning, and late layers drive language-specific generation. Guided by these insights, we introduce selective fine-tuning of only the final layers responsible for language control. On Qwen-3-32B and Bloom-7.1B, this method achieves over 98 percent language consistency across six languages while fine-tuning only 3-5 percent of parameters, without sacrificing task accuracy. Importantly, this result is nearly identical to that of full-scope fine-tuning (for example, above 98 percent language consistency for both methods across all prompt scenarios) but uses a fraction of the computational resources. To the best of our knowledge, this is the first approach to leverage layer-localization of language control for efficient multilingual adaptation.
Authors: Fang Li
Abstract: Despite the ubiquity of tabular data in high-stakes domains, traditional deep learning architectures often struggle to match the performance of gradient-boosted decision trees while maintaining scientific interpretability. Standard neural networks typically treat features as independent entities, failing to exploit the inherent manifold structural dependencies that define tabular distributions. We propose Structural Compositional Function Networks (StructuralCFN), a novel architecture that imposes a Relation-Aware Inductive Bias via a differentiable structural prior. StructuralCFN explicitly models each feature as a mathematical composition of its counterparts through Differentiable Adaptive Gating, which automatically discovers the optimal activation physics (e.g., attention-style filtering vs. inhibitory polarity) for each relationship. Our framework enables Structured Knowledge Integration, allowing domain-specific relational priors to be injected directly into the architecture to guide discovery. We evaluate StructuralCFN across a rigorous 10-fold cross-validation suite on 18 benchmarks, demonstrating statistically significant improvements (p < 0.05) on scientific and clinical datasets (e.g., Blood Transfusion, Ozone, WDBC). Furthermore, StructuralCFN provides Intrinsic Symbolic Interpretability: it recovers the governing "laws" of the data manifold as human-readable mathematical expressions while maintaining a compact parameter footprint (300--2,500 parameters) that is over an order of magnitude (10x--20x) smaller than standard deep baselines.
Authors: Shih-Hsuan Chiu, Ming-Syan Chen
Abstract: Personalized virtual assistants powered by large language models (LLMs) on edge devices are attracting growing attention, with Retrieval-Augmented Generation (RAG) emerging as a key method for personalization by retrieving relevant profile data and generating tailored responses. However, deploying RAG on edge devices faces efficiency hurdles due to the rapid growth of profile data, such as user-LLM interactions and recent updates. While Computing-in-Memory (CiM) architectures mitigate this bottleneck by eliminating data movement between memory and processing units via in-situ operations, they are susceptible to environmental noise that can degrade retrieval precision. This poses a critical issue in dynamic, multi-domain edge-based scenarios (e.g., travel, medicine, and law) where both accuracy and adaptability are paramount. To address these challenges, we propose Task-Oriented Noise-resilient Embedding Learning (TONEL), a framework that improves noise robustness and domain adaptability for RAG in noisy edge environments. TONEL employs a noise-aware projection model to learn task-specific embeddings compatible with CiM hardware constraints, enabling accurate retrieval under noisy conditions. Extensive experiments conducted on personalization benchmarks demonstrate the effectiveness and practicality of our methods relative to strong baselines, especially in task-specific noisy scenarios.
Authors: Gautham Vinod, Bruce Coburn, Siddeshwar Raghavan, Jiangpeng He, Fengqing Zhu
Abstract: The rise of chronic diseases related to diet, such as obesity and diabetes, emphasizes the need for accurate monitoring of food intake. While AI-driven dietary assessment has made strides in recent years, the ill-posed nature of recovering size (portion) information from monocular images for accurate estimation of ``how much did you eat?'' is a pressing challenge. Some 3D reconstruction methods have achieved impressive geometric reconstruction but fail to recover the crucial real-world scale of the reconstructed object, limiting its usage in precision nutrition. In this paper, we bridge the gap between 3D computer vision and digital health by proposing a method that recovers a true-to-scale 3D reconstructed object from a monocular image. Our approach leverages rich visual features extracted from models trained on large-scale datasets to estimate the scale of the reconstructed object. This learned scale enables us to convert single-view 3D reconstructions into true-to-life, physically meaningful models. Extensive experiments and ablation studies on two publicly available datasets show that our method consistently outperforms existing techniques, achieving nearly a 30% reduction in mean absolute volume-estimation error, showcasing its potential to enhance the domain of precision nutrition. Code: https://gitlab.com/viper-purdue/size-matters
Authors: Vikash Singh, Darion Cassel, Nathaniel Weir, Nick Feng, Sam Bayless
Abstract: Despite the syntactic fluency of Large Language Models (LLMs), ensuring their logical correctness in high-stakes domains remains a fundamental challenge. We present a neurosymbolic framework that combines LLMs with SMT solvers to produce verification-guided answers through iterative refinement. Our approach decomposes LLM outputs into atomic claims, autoformalizes them into first-order logic, and verifies their logical consistency using automated theorem proving. We introduce three key innovations: (1) multi-model consensus via formal semantic equivalence checking to ensure logic-level alignment between candidates, eliminating the syntactic bias of surface-form metrics, (2) semantic routing that directs different claim types to appropriate verification strategies: symbolic solvers for logical claims and LLM ensembles for commonsense reasoning, and (3) precise logical error localization via Minimal Correction Subsets (MCS), which pinpoint the exact subset of claims to revise, transforming binary failure signals into actionable feedback. Our framework classifies claims by their logical status and aggregates multiple verification signals into a unified score with variance-based penalty. The system iteratively refines answers using structured feedback until acceptance criteria are met or convergence is achieved. This hybrid approach delivers formal guarantees where possible and consensus verification elsewhere, advancing trustworthy AI. With the GPT-OSS-120B model, VERGE demonstrates an average performance uplift of 18.7% at convergence across a set of reasoning benchmarks compared to single-pass approaches.
Authors: Atik Faysal, Mohammad Rostami, Reihaneh Gh. Roshan, Nikhil Muralidhar, Huaxia Wang
Abstract: We address the challenge of training Vision Transformers (ViTs) when labeled data is scarce but unlabeled data is abundant. We propose Semi-Supervised Masked Autoencoder (SSMAE), a framework that jointly optimizes masked image reconstruction and classification using both unlabeled and labeled samples with dynamically selected pseudo-labels. SSMAE introduces a validation-driven gating mechanism that activates pseudo-labeling only after the model achieves reliable, high-confidence predictions that are consistent across both weakly and strongly augmented views of the same image, reducing confirmation bias. On CIFAR-10 and CIFAR-100, SSMAE consistently outperforms supervised ViT and fine-tuned MAE, with the largest gains in low-label regimes (+9.24% over ViT on CIFAR-10 with 10% labels). Our results demonstrate that when pseudo-labels are introduced is as important as how they are generated for data-efficient transformer training. Codes are available at https://github.com/atik666/ssmae.
Authors: Lee Xiong, Zhirong Chen, Rahul Mayuranath, Shangran Qiu, Arda Ozdemir, Lu Li, Yang Hu, Dave Li, Jingtao Ren, Howard Cheng, Fabian Souto Herrera, Ahmed Agiza, Baruch Epshtein, Anuj Aggarwal, Julia Ulziisaikhan, Chao Wang, Dinesh Ramasamy, Parshva Doshi, Sri Reddy, Arnold Overwijk
Abstract: We present LLaTTE (LLM-Style Latent Transformers for Temporal Events), a scalable transformer architecture for production ads recommendation. Through systematic experiments, we demonstrate that sequence modeling in recommendation systems follows predictable power-law scaling similar to LLMs. Crucially, we find that semantic features bend the scaling curve: they are a prerequisite for scaling, enabling the model to effectively utilize the capacity of deeper and longer architectures. To realize the benefits of continued scaling under strict latency constraints, we introduce a two-stage architecture that offloads the heavy computation of large, long-context models to an asynchronous upstream user model. We demonstrate that upstream improvements transfer predictably to downstream ranking tasks. Deployed as the largest user model at Meta, this multi-stage framework drives a 4.3\% conversion uplift on Facebook Feed and Reels with minimal serving overhead, establishing a practical blueprint for harnessing scaling laws in industrial recommender systems.
Authors: Buddhika Nettasinghe, Kang Zhao
Abstract: Humans and large language models (LLMs) now co-produce and co-consume the web's shared knowledge archives. Such human-AI collective knowledge ecosystems contain feedback loops with both benefits (e.g., faster growth, easier learning) and systemic risks (e.g., quality dilution, skill reduction, model collapse). To understand such phenomena, we propose a minimal, interpretable dynamical model of the co-evolution of archive size, archive quality, model (LLM) skill, aggregate human skill, and query volume. The model captures two content inflows (human, LLM) controlled by a gate on LLM-content admissions, two learning pathways for humans (archive study vs. LLM assistance), and two LLM-training modalities (corpus-driven scaling vs. learning from human feedback). Through numerical experiments, we identify different growth regimes (e.g., healthy growth, inverted flow, inverted learning, oscillations), and show how platform and policy levers (gate strictness, LLM training, human learning pathways) shift the system across regime boundaries. Two domain configurations (PubMed, GitHub and Copilot) illustrate contrasting steady states under different growth rates and moderation norms. We also fit the model to Wikipedia's knowledge flow during pre-ChatGPT and post-ChatGPT eras separately. We find a rise in LLM additions with a concurrent decline in human inflow, consistent with a regime identified by the model. Our model and analysis yield actionable insights for sustainable growth of human-AI collective knowledge on the Web.
Authors: Jeremy Foote, Deepak Kumar, Bedadyuti Jha, Ryan Funkhouser, Loizos Bitsikokos, Hitesh Goel, Hsuen-Chi Chiu
Abstract: Generative AI chatbots have proven surprisingly effective at persuading people to change their beliefs and attitudes in lab settings. However, the practical implications of these findings are not yet clear. In this work, we explore the impact of rehabilitative conversations with generative AI chatbots on users who share toxic content online. Toxic behaviors -- like insults or threats of violence, are widespread in online communities. Strategies to deal with toxic behavior are typically punitive, such as removing content or banning users. Rehabilitative approaches are rarely attempted, in part due to the emotional and psychological cost of engaging with aggressive users. In collaboration with seven large Reddit communities, we conducted a large-scale field experiment (N=893) to invite people who had recently posted toxic content to participate in conversations with AI chatbots. A qualitative analysis of the conversations shows that many participants engaged in good faith and even expressed remorse or a desire to change. However, we did not observe a significant change in toxic behavior in the following month compared to a control group. We discuss possible explanations for our findings, as well as theoretical and practical implications based on our results.
Authors: Darshan Deshpande, Anand Kannappan, Rebecca Qian
Abstract: Recent advances in reinforcement learning for code generation have made robust environments essential to prevent reward hacking. As LLMs increasingly serve as evaluators in code-based RL, their ability to detect reward hacking remains understudied. In this paper, we propose a novel taxonomy of reward exploits spanning across 54 categories and introduce TRACE (Testing Reward Anomalies in Code Environments), a synthetically curated and human-verified benchmark containing 517 testing trajectories. Unlike prior work that evaluates reward hack detection in isolated classification scenarios, we contrast these evaluations with a more realistic, contrastive anomaly detection setup on TRACE. Our experiments reveal that models capture reward hacks more effectively in contrastive settings than in isolated classification settings, with GPT-5.2 with highest reasoning mode achieving the best detection rate at 63%, up from 45% in isolated settings on TRACE. Building on this insight, we demonstrate that state-of-the-art models struggle significantly more with semantically contextualized reward hacks compared to syntactically contextualized ones. We further conduct qualitative analyses of model behaviors, as well as ablation studies showing that the ratio of benign to hacked trajectories and analysis cluster sizes substantially impact detection performance. We release the benchmark and evaluation harness to enable the community to expand TRACE and evaluate their models.
Authors: Emanuele Del Sozzo, Martin Fleming, Kenneth Flamm, Neil Thompson
Abstract: Graphics Processing Units (GPUs) are the state-of-the-art architecture for essential tasks, ranging from rendering 2D/3D graphics to accelerating workloads in supercomputing centers and, of course, Artificial Intelligence (AI). As GPUs continue improving to satisfy ever-increasing performance demands, analyzing past and current progress becomes paramount in determining future constraints on scientific research. This is particularly compelling in the AI domain, where rapid technological advancements and fierce global competition have led the United States to recently implement export control regulations limiting international access to advanced AI chips. For this reason, this paper studies technical progress in NVIDIA datacenter GPUs released from the mid-2000s until today. Specifically, we compile a comprehensive dataset of datacenter NVIDIA GPUs comprising several features, ranging from computational performance to release price. Then, we examine trends in main GPU features and estimate progress indicators for per-memory bandwidth, per-dollar, and per-watt increase rates. Our main results identify doubling times of 1.44 and 1.69 years for FP16 and FP32 operations (without accounting for sparsity benefits), while FP64 doubling times range from 2.06 to 3.79 years. Off-chip memory size and bandwidth grew at slower rates than computing performance, doubling every 3.32 to 3.53 years. The release prices of datacenter GPUs have roughly doubled every 5.1 years, while their power consumption has approximately doubled every 16 years. Finally, we quantify the potential implications of current U.S. export control regulations in terms of the potential performance gaps that would result if implementation were assumed to be complete and successful. We find that recently proposed changes to export controls would shrink the potential performance gap from 23.6x to 3.54x.
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: Abha Jha, Akanksha Mahajan, Ashwath Vaithinathan Aravindan, Praveen Saravanan, Sai Sailaja Policharla, Sonal Chaturbhuj Gehlot
Abstract: Large Language Models (LLMs) often produce hallucinated or unverifiable content, undermining their reliability in factual domains. This work investigates Reinforcement Learning with Verifiable Rewards (RLVR) as a training paradigm that explicitly rewards abstention ("I don't know") alongside correctness to promote intellectual humility. We fine-tune and evaluate Granite-3.3-2B-Instruct and Qwen-3-4B-Instruct on the MedMCQA and Hendrycks Math benchmarks using a ternary reward structure ($-1$, r_abs, 1) under varying abstention reward structures. We further study the effect of combining RLVR with supervised fine-tuning strategies that teach abstention prior to reinforcement learning. Our results show that moderate abstention rewards (r_abs $\approx -0.25$ to 0.3) consistently reduce incorrect responses without severe accuracy degradation on multiple-choice tasks, with larger models exhibiting greater robustness to abstention incentives. On open-ended question answering, we observe limitations due to insufficient exploration, which can be partially mitigated through supervised abstention training. Overall, these findings demonstrate the feasibility and flexibility of verifiable reward design as a practical approach for hallucination mitigation in language models. Reproducible code for our abstention training framework is available here https://github.com/Mystic-Slice/rl-abstention.
Authors: Akif Islam, Sujan Kumar Roy, Md. Ekramul Hamid
Abstract: Sentiment analysis for the Bengali language has attracted increasing research interest in recent years. However, progress remains constrained by the scarcity of large-scale and diverse annotated datasets. Although several Bengali sentiment and hate speech datasets are publicly available, most are limited in size or confined to a single domain, such as social media comments. Consequently, these resources are often insufficient for training modern deep learning based models, which require large volumes of heterogeneous data to learn robust and generalizable representations. In this work, we introduce BengaliSent140, a large-scale Bengali binary sentiment dataset constructed by consolidating seven existing Bengali text datasets into a unified corpus. To ensure consistency across sources, heterogeneous annotation schemes are systematically harmonized into a binary sentiment formulation with two classes: Not Hate (0) and Hate (1). The resulting dataset comprises 139,792 unique text samples, including 68,548 hate and 71,244 not-hate instances, yielding a relatively balanced class distribution. By integrating data from multiple sources and domains, BengaliSent140 offers broader linguistic and contextual coverage than existing Bengali sentiment datasets and provides a strong foundation for training and benchmarking deep learning models. Baseline experimental results are also reported to demonstrate the practical usability of the dataset. The dataset is publicly available at https://www.kaggle.com/datasets/akifislam/bengalisent140/
URLs: https://www.kaggle.com/datasets/akifislam/bengalisent140/
Authors: Deep Shah, Sanket Badhe, Nehal Kathrotia
Abstract: Designing an embedding retrieval system requires navigating a complex design space of conflicting trade-offs between efficiency and effectiveness. This work structures these decisions as a vertical traversal of the system design stack. We begin with the Representation Layer by examining how loss functions and architectures, specifically Bi-encoders and Cross-encoders, define semantic relevance and geometric projection. Next, we analyze the Granularity Layer and evaluate how segmentation strategies like Atomic and Hierarchical chunking mitigate information bottlenecks in long-context documents. Moving to the Orchestration Layer, we discuss methods that transcend the single-vector paradigm, including hierarchical retrieval, agentic decomposition, and multi-stage reranking pipelines to resolve capacity limitations. Finally, we address the Robustness Layer by identifying architectural mitigations for domain generalization failures, lexical blind spots, and the silent degradation of retrieval quality due to temporal drift. By categorizing these limitations and design choices, we provide a comprehensive framework for practitioners to optimize the efficiency-effectiveness frontier in modern neural search systems.
Authors: Nattavudh Powdthavee, Sandra J. Geiger
Abstract: Although most people support climate action, widespread underestimation of others' support stalls individual and systemic changes. In this preregistered experiment, we test whether large language models (LLMs) can reliably predict these perception gaps worldwide. Using country-level indicators and public opinion data from 125 countries, we benchmark four state-of-the-art LLMs against Gallup World Poll 2021/22 data and statistical regressions. LLMs, particularly Claude, accurately capture public perceptions of others' willingness to contribute financially to climate action (MAE approximately 5 p.p.; r = .77), comparable to statistical models, though performance declines in less digitally connected, lower-GDP countries. Controlled tests show that LLMs capture the key psychological process - social projection with a systematic downward bias - and rely on structured reasoning rather than memorized values. Overall, LLMs provide a rapid tool for assessing perception gaps in climate action, serving as an alternative to costly surveys in resource-rich countries and as a complement in underrepresented populations.
Authors: Jim Maar, Denis Paperno, Callum Stuart McDougall, Neel Nanda
Abstract: Prior work suggests that language models, while trained on next token prediction, show implicit planning behavior: they may select the next token in preparation to a predicted future token, such as a likely rhyming word, as supported by a prior qualitative study of Claude 3.5 Haiku using a cross-layer transcoder. We propose much simpler techniques for assessing implicit planning in language models. With case studies on rhyme poetry generation and question answering, we demonstrate that our methodology easily scales to many models. Across models, we find that the generated rhyme (e.g. "-ight") or answer to a question ("whale") can be manipulated by steering at the end of the preceding line with a vector, affecting the generation of intermediate tokens leading up to the rhyme or answer word. We show that implicit planning is a universal mechanism, present in smaller models than previously thought, starting from 1B parameters. Our methodology offers a widely applicable direct way to study implicit planning abilities of LLMs. More broadly, understanding planning abilities of language models can inform decisions in AI safety and control.
Authors: Alexander Benanti, Xi Han, Hong Qin
Abstract: Numerical techniques for solving partial differential equations (PDEs) are integral for many fields across science and engineering. Such techniques usually involve solving large, sparse linear systems, where preconditioning methods are critical. In recent years, neural methods, particularly graph neural networks (GNNs), have demonstrated their potential through accelerated convergence. Nonetheless, to extract connective structures, existing techniques aggregate discretized system matrices into graphs, and suffer from rank inflation and a suboptimal convergence rate. In this paper, we articulate NeuraLSP, a novel neural preconditioner combined with a novel loss metric that leverages the left singular subspace of the system matrix's near-nullspace vectors. By compressing spectral information into a fixed low-rank operator, our method exhibits both theoretical guarantees and empirical robustness to rank inflation, affording up to a 53% speedup. Besides the theoretical guarantees for our newly-formulated loss function, our comprehensive experimental results across diverse families of PDEs also substantiate the aforementioned theoretical advances.
Authors: Chen Cheng, Ang Li
Abstract: Out-of-distribution (OOD) generalization remains a fundamental challenge in real-world classification, where test distributions often differ substantially from training data. Most existing approaches pursue domain-invariant representations, implicitly assuming that invariance implies reliability. However, features that are invariant across domains are not necessarily causally effective for prediction. In this work, we revisit OOD classification from a causal perspective and propose to evaluate learned representations based on their necessity and sufficiency under distribution shift. We introduce an explicit segment-level framework that directly measures causal effectiveness across domains, providing a more faithful criterion than invariance alone. Experiments on multi-domain benchmarks demonstrate consistent improvements in OOD performance, particularly under challenging domain shifts, highlighting the value of causal evaluation for robust generalization.
Authors: Zhipeng Zhang, Wenting Ma, Kai Li, Meng Guo, Lei Yang, Wei Yu, Hongji Cui, Yichen Zhang, Mo Zhang, Jinzhe Lin, Zhenjie Yao
Abstract: Robust reinforcement learning methods typically focus on suppressing unreliable experiences or corrupted rewards, but they lack the ability to reason about the reliability of their own learning process. As a result, such methods often either overreact to noise by becoming overly conservative or fail catastrophically when uncertainty accumulates. In this work, we propose a meta-cognitive reinforcement learning framework that enables an agent to assess, regulate, and recover its learning behavior based on internally estimated reliability signals. The proposed method introduces a meta-trust variable driven by Value Prediction Error Stability (VPES), which modulates learning dynamics via fail-safe regulation and gradual trust recovery. Experiments on continuous-control benchmarks with reward corruption demonstrate that recovery-enabled meta-cognitive control achieves higher average returns and significantly reduces late-stage training failures compared to strong robustness baselines.
Authors: Zichao Yu, Ming Li, Wenyi Zhang, Difan Zou, Weiguo Gao
Abstract: Inferring physical fields from sparse observations while strictly satisfying partial differential equations (PDEs) is a fundamental challenge in computational physics. Recently, deep generative models offer powerful data-driven priors for such inverse problems, yet existing methods struggle to enforce hard physical constraints without costly retraining or disrupting the learned generative prior. Consequently, there is a critical need for a sampling mechanism that can reconcile strict physical consistency and observational fidelity with the statistical structure of the pre-trained prior. To this end, we present ProFlow, a proximal guidance framework for zero-shot physics-consistent sampling, defined as inferring solutions from sparse observations using a fixed generative prior without task-specific retraining. The algorithm employs a rigorous two-step scheme that alternates between: (\romannumeral1) a terminal optimization step, which projects the flow prediction onto the intersection of the physically and observationally consistent sets via proximal minimization; and (\romannumeral2) an interpolation step, which maps the refined state back to the generative trajectory to maintain consistency with the learned flow probability path. This procedure admits a Bayesian interpretation as a sequence of local maximum a posteriori (MAP) updates. Comprehensive benchmarks on Poisson, Helmholtz, Darcy, and viscous Burgers' equations demonstrate that ProFlow achieves superior physical and observational consistency, as well as more accurate distributional statistics, compared to state-of-the-art diffusion- and flow-based baselines.
Authors: Ibne Farabi Shihab, Sanjeda Akter, Anuj Sharma
Abstract: We study black-box optimization of Lipschitz functions under noisy evaluations. Existing adaptive discretization methods implicitly avoid suboptimal regions but do not provide explicit certificates of optimality or measurable progress guarantees. We introduce \textbf{Certificate-Guided Pruning (CGP)}, which maintains an explicit \emph{active set} $A_t$ of potentially optimal points via confidence-adjusted Lipschitz envelopes. Any point outside $A_t$ is certifiably suboptimal with high probability, and under a margin condition with near-optimality dimension $\alpha$, we prove $\Vol(A_t)$ shrinks at a controlled rate yielding sample complexity $\tildeO(\varepsilon^{-(2+\alpha)})$. We develop three extensions: CGP-Adaptive learns $L$ online with $O(\log T)$ overhead; CGP-TR scales to $d > 50$ via trust regions with local certificates; and CGP-Hybrid switches to GP refinement when local smoothness is detected. Experiments on 12 benchmarks ($d \in [2, 100]$) show CGP variants match or exceed strong baselines while providing principled stopping criteria via certificate volume.
Authors: Qihang Yu, Kairui Fu, Zhaocheng Du, Yuxuan Si, Kaiyuan Li, Weihao Zhao, Zhicheng Zhang, Jieming Zhu, Quanyu Dai, Zhenhua Dong, Shengyu Zhang, Kun Kuang, Fei Wu
Abstract: The scaling law, which indicates that model performance improves with increasing dataset and model capacity, has fueled a growing trend in expanding recommendation models in both industry and academia. However, the advent of large-scale recommenders also brings significantly higher computational costs, particularly under the long-sequence dependencies inherent in the user intent of recommendation systems. Current approaches often rely on pre-storing the intermediate states of the past behavior for each user, thereby reducing the quadratic re-computation cost for the following requests. Despite their effectiveness, these methods often treat memory merely as a medium for acceleration, without adequately considering the space overhead it introduces. This presents a critical challenge in real-world recommendation systems with billions of users, each of whom might initiate thousands of interactions and require massive memory for state storage. Fortunately, there have been several memory management strategies examined for compression in LLM, while most have not been evaluated on the recommendation task. To mitigate this gap, we introduce MALLOC, a comprehensive benchmark for memory-aware long sequence compression. MALLOC presents a comprehensive investigation and systematic classification of memory management techniques applicable to large sequential recommendations. These techniques are integrated into state-of-the-art recommenders, enabling a reproducible and accessible evaluation platform. Through extensive experiments across accuracy, efficiency, and complexity, we demonstrate the holistic reliability of MALLOC in advancing large-scale recommendation. Code is available at https://anonymous.4open.science/r/MALLOC.
Authors: Judy Hanwen Shen, Alex Tamkin
Abstract: AI assistance produces significant productivity gains across professional domains, particularly for novice workers. Yet how this assistance affects the development of skills required to effectively supervise AI remains unclear. Novice workers who rely heavily on AI to complete unfamiliar tasks may compromise their own skill acquisition in the process. We conduct randomized experiments to study how developers gained mastery of a new asynchronous programming library with and without the assistance of AI. We find that AI use impairs conceptual understanding, code reading, and debugging abilities, without delivering significant efficiency gains on average. Participants who fully delegated coding tasks showed some productivity improvements, but at the cost of learning the library. We identify six distinct AI interaction patterns, three of which involve cognitive engagement and preserve learning outcomes even when participants receive AI assistance. Our findings suggest that AI-enhanced productivity is not a shortcut to competence and AI assistance should be carefully adopted into workflows to preserve skill formation -- particularly in safety-critical domains.
Authors: Hari Krishna Sahoo, Mudit Gaur, Vaneet Aggarwal
Abstract: Recently, flow-based generative models have shown superior efficiency compared to diffusion models. In this paper, we study rectified flow models, which constrain transport trajectories to be linear from the base distribution to the data distribution. This structural restriction greatly accelerates sampling, often enabling high-quality generation with a single Euler step. Under standard assumptions on the neural network classes used to parameterize the velocity field and data distribution, we prove that rectified flows achieve sample complexity $\tilde{O}(\varepsilon^{-2})$. This improves on the best known $O(\varepsilon^{-4})$ bounds for flow matching model and matches the optimal rate for mean estimation. Our analysis exploits the particular structure of rectified flows: because the model is trained with a squared loss along linear paths, the associated hypothesis class admits a sharply controlled localized Rademacher complexity. This yields the improved, order-optimal sample complexity and provides a theoretical explanation for the strong empirical performance of rectified flow models.
Authors: Si Chen, Le Huy Khiem, Annalisa Szymanski, Ronald Metoyer, Ting Hua, Nitesh V. Chawla
Abstract: Open-ended question answering (QA) evaluates a model's ability to perform contextualized reasoning beyond factual recall. This challenge is especially acute in practice-based domains, where knowledge is procedural and grounded in professional judgment, while most existing LLM benchmarks depend on pre-existing human exam datasets that are often unavailable in such settings. We introduce a framework for automated benchmark generation from expert-authored guidelines informed by Bloom's Taxonomy. It converts expert practices into implicit violation-based scenarios and expands them into auto-graded multiple-choice questions (MCQs) and multi-turn dialogues across four cognitive levels, enabling deterministic, reproducible, and scalable evaluation. Applied to three applied domains: teaching, dietetics, and caregiving, we find differences between model and human-like reasoning: LLMs sometimes perform relatively better on higher-order reasoning (Analyze) but fail more frequently on lower-level items (Remember). We produce large-scale, psychometrically informed benchmarks that surface these non-intuitive model behaviors and enable evaluation of contextualized reasoning in real-world settings.
Authors: Long Van Tran, Truyen Tran, Phuoc Nguyen
Abstract: Recent advances in stochastic differential equations (SDEs) have enabled robust modeling of real-world dynamical processes across diverse domains, such as finance, health, and systems biology. However, parameter estimation for SDEs typically relies on accurately timestamped observational sequences. When temporal ordering information is corrupted, missing, or deliberately hidden (e.g., for privacy), existing estimation methods often fail. In this paper, we investigate the conditions under which temporal order can be recovered and introduce a novel framework that simultaneously reconstructs temporal information and estimates SDE parameters. Our approach exploits asymmetries between forward and backward processes, deriving a score-matching criterion to infer the correct temporal order between pairs of observations. We then recover the total order via a sorting procedure and estimate SDE parameters from the reconstructed sequence using maximum likelihood. Finally, we conduct extensive experiments on synthetic and real-world datasets to demonstrate the effectiveness of our method, extending parameter estimation to settings with missing temporal order and broadening applicability in sensitive domains.
Authors: Holly Trikilis, Pasindu Marasinghe, Fariza Rashid, Suranga Seneviratne
Abstract: Phishing continues to be one of the most prevalent attack vectors, making accurate classification of phishing URLs essential. Recently, large language models (LLMs) have demonstrated promising results in phishing URL detection. However, their reasoning capabilities that enabled such performance remain underexplored. To this end, in this paper, we propose a Least-to-Most prompting framework for phishing URL detection. In particular, we introduce an "answer sensitivity" mechanism that guides Least-to-Most's iterative approach to enhance reasoning and yield higher prediction accuracy. We evaluate our framework using three URL datasets and four state-of-the-art LLMs, comparing against a one-shot approach and a supervised model. We demonstrate that our framework outperforms the one-shot baseline while achieving performance comparable to that of the supervised model, despite requiring significantly less training data. Furthermore, our in-depth analysis highlights how the iterative reasoning enabled by Least-to-Most, and reinforced by our answer sensitivity mechanism, drives these performance gains. Overall, we show that this simple yet powerful prompting strategy consistently outperforms both one-shot and supervised approaches, despite requiring minimal training or few-shot guidance. Our experimental setup can be found in our Github repository github.sydney.edu.au/htri0928/least-to-most-phishing-detection.
Authors: Tianwei Lin, Zuyi Zhou, Xinda Zhao, Chenke Wang, Xiaohong Li, Yu Chen, Chuanrui Hu, Jian Pei, Yafeng Deng
Abstract: Long-context LLM agents must access the right evidence from large environments and use it faithfully. However, the popular Needle-in-a-Haystack (NIAH) evaluation mostly measures benign span localization. The needle is near-unique, and the haystack is largely irrelevant. We introduce EverMemBench-S (EMB-S), an adversarial NIAH-style benchmark built on a 326M-token MemoryBank. While the full MemoryBank spans 326M tokens for retrieval-based (RAG) evaluation, we evaluate native long-context models only at scales that fit within each model's context window (up to 1M tokens in this work) to ensure a fair comparison. EMB-S pairs queries with collision-tested near-miss hard negatives and gold evidence sets spanning one or more documents, validated via human screening and LLM verification. We also propose a decoupled diagnostic protocol that reports evidence access (document-ID localization) separately from end-to-end QA quality under full-context prompting. This enables consistent diagnosis for both native long-context prompting and retrieval pipelines. Across a reference-corpus ladder from domain-isolated 64K contexts to a globally shared 326M-token environment, we observe a clear reality gap. Systems that saturate benign NIAH degrade sharply in evidence access under semantic interference. These results indicate that semantic discrimination, not context length alone, is the dominant bottleneck for long-context memory at scale.
Authors: Daojun Liang, Qi Li, Yinglong Wang, Jing Chen, Hu Zhang, Xiaoxiao Cui, Qizheng Wang, Shuo Li
Abstract: Time series forecasting has long been dominated by advances in model architecture, with recent progress driven by deep learning and hybrid statistical techniques. However, as forecasting models approach diminishing returns in accuracy, a critical yet underexplored opportunity emerges: the strategic use of post-processing. In this paper, we address the last-mile gap in time-series forecasting, which is to improve accuracy and uncertainty without retraining or modifying a deployed backbone. We propose $\delta$-Adapter, a lightweight, architecture-agnostic way to boost deployed time series forecasters without retraining. $\delta$-Adapter learns tiny, bounded modules at two interfaces: input nudging (soft edits to covariates) and output residual correction. We provide local descent guarantees, $O(\delta)$ drift bounds, and compositional stability for combined adapters. Meanwhile, it can act as a feature selector by learning a sparse, horizon-aware mask over inputs to select important features, thereby improving interpretability. In addition, it can also be used as a distribution calibrator to measure uncertainty. Thus, we introduce a Quantile Calibrator and a Conformal Corrector that together deliver calibrated, personalized intervals with finite-sample coverage. Our experiments across diverse backbones and datasets show that $\delta$-Adapter improves accuracy and calibration with negligible compute and no interface changes.
Authors: Yuxuan Bao, Jan Zajac, Megan Powers, Venkat Raman, J. Nathan Kutz
Abstract: Bridging the sim2real gap between computationally inexpensive models and complex physical systems remains a central challenge in machine learning applications to engineering problems, particularly in multi-scale settings where reduced-order models typically capture only dominant dynamics. In this work, we present Cheap2Rich, a multi-scale data assimilation framework that reconstructs high-fidelity state spaces from sparse sensor histories by combining a fast low-fidelity prior with learned, interpretable discrepancy corrections. We demonstrate the performance on rotating detonation engines (RDEs), a challenging class of systems that couple detonation-front propagation with injector-driven unsteadiness, mixing, and stiff chemistry across disparate scales. Our approach successfully reconstructs high-fidelity RDE states from sparse measurements while isolating physically meaningful discrepancy dynamics associated with injector-driven effects. The results highlight a general multi-fidelity framework for data assimilation and system identification in complex multi-scale systems, enabling rapid design exploration and real-time monitoring and control while providing interpretable discrepancy dynamics. Code for this project is is available at: github.com/kro0l1k/Cheap2Rich.
Authors: Tianyi Alex Qiu, Micah Carroll, Cameron Allen
Abstract: The evaluation and post-training of large language models (LLMs) rely on supervision, but strong supervision for difficult tasks is often unavailable, especially when evaluating frontier models. In such cases, models are demonstrated to exploit evaluations built on such imperfect supervision, leading to deceptive results. However, underutilized in LLM research, a wealth of mechanism design research focuses on game-theoretic incentive compatibility, i.e., eliciting honest and informative answers with weak supervision. Drawing from this literature, we introduce the peer prediction method for model evaluation and post-training. It rewards honest and informative answers over deceptive and uninformative ones, using a metric based on mutual predictability and without requiring ground truth labels. We demonstrate the method's effectiveness and resistance to deception, with both theoretical guarantees and empirical validation on models with up to 405B parameters. We show that training an 8B model with peer prediction-based reward recovers most of the drop in truthfulness due to prior malicious finetuning, even when the reward is produced by a 0.135B language model with no finetuning. On the evaluation front, in contrast to LLM-as-a-Judge which requires strong and trusted judges, we discover an inverse scaling property in peer prediction, where, surprisingly, resistance to deception is strengthened as the capability gap between the experts and participants widens, enabling reliable evaluation of strong models with weak supervision. In particular, LLM-as-a-Judge become worse than random guess when facing deceptive models 5-20x the judge's size, while peer prediction thrives when such gaps are large, including in cases with over 100x size difference.
Authors: Jialuo He, Huangxun Chen
Abstract: Sharpness-Aware Minimization (SAM) has recently emerged as an effective technique for improving DNN robustness to input variations. However, its interplay with the compactness requirements of on-device DNN deployments remains less explored. Simply pruning a SAM-trained model can undermine robustness, since flatness in the continuous parameter space does not necessarily translate to robustness under the discrete structural changes induced by pruning. Conversely, applying SAM after pruning may be fundamentally constrained by architectural limitations imposed by an early, robustness-agnostic pruning pattern. To address this gap, we propose Compression-aware ShArpness Minimization (C-SAM), a framework that shifts sharpness-aware learning from parameter perturbations to mask perturbations. By explicitly perturbing pruning masks during training, C-SAM promotes a flatter loss landscape with respect to model structure, enabling the discovery of pruning patterns that simultaneously optimize model compactness and robustness to input variations. Extensive experiments on CelebA-HQ, Flowers-102, and CIFAR-10-C across ResNet-18, GoogLeNet, and MobileNet-V2 show that C-SAM consistently achieves higher certified robustness than strong baselines, with improvements of up to 42%, while maintaining task accuracy comparable to the corresponding unpruned models.
Authors: Sungjae Lee, Junhan Jeong, Yeonjoo Hong, Kwang In Kim
Abstract: Estimating object mass from visual input is challenging because mass depends jointly on geometric volume and material-dependent density, neither of which is directly observable from RGB appearance. Consequently, mass prediction from pixels is ill-posed and therefore benefits from physically meaningful representations to constrain the space of plausible solutions. We propose a physically structured framework for single-image mass estimation that addresses this ambiguity by aligning visual cues with the physical factors governing mass. From a single RGB image, we recover object-centric three-dimensional geometry via monocular depth estimation to inform volume and extract coarse material semantics using a vision-language model to guide density-related reasoning. These geometry, semantic, and appearance representations are fused through an instance-adaptive gating mechanism, and two physically guided latent factors (volume- and density-related) are predicted through separate regression heads under mass-only supervision. Experiments on image2mass and ABO-500 show that the proposed method consistently outperforms state-of-the-art methods.
Authors: Genyuan Zhang, Zihao Wang, Zhifan Gao, Lei Xu, Zhen Zhou, Haijun Yu, Jianjia Zhang, Xiujian Liu, Weiwei Zhang, Shaoyu Wang, Huazhu Fu, Fenglin Liu, Weiwen Wu
Abstract: The application of iodinated contrast media (ICM) improves the sensitivity and specificity of computed tomography (CT) for a wide range of clinical indications. However, overdose of ICM can cause problems such as kidney damage and life-threatening allergic reactions. Deep learning methods can generate CT images of normal-dose ICM from low-dose ICM, reducing the required dose while maintaining diagnostic power. However, existing methods are difficult to realize accurate enhancement with incompletely paired images, mainly because of the limited ability of the model to recognize specific structures. To overcome this limitation, we propose a Structure-constrained Language-informed Diffusion Model (SLDM), a unified medical generation model that integrates structural synergy and spatial intelligence. First, the structural prior information of the image is effectively extracted to constrain the model inference process, thus ensuring structural consistency in the enhancement process. Subsequently, semantic supervision strategy with spatial intelligence is introduced, which integrates the functions of visual perception and spatial reasoning, thus prompting the model to achieve accurate enhancement. Finally, the subtraction angiography enhancement module is applied, which serves to improve the contrast of the ICM agent region to suitable interval for observation. Qualitative analysis of visual comparison and quantitative results of several metrics demonstrate the effectiveness of our method in angiographic reconstruction for low-dose contrast medium CT angiography.
Authors: Jiahuan Yu, Mingtao Hu, Zichao Lin, Minjia Zhang
Abstract: Large Language Model (LLM) serving faces a fundamental tension between stringent latency Service Level Objectives (SLOs) and limited GPU memory capacity. When high request rates exhaust the KV cache budget, existing LLM inference systems often suffer severe head-of-line (HOL) blocking. While prior work explored PCIe-based offloading, these approaches cannot sustain responsiveness under high request rates, often failing to meet tight Time-To-First-Token (TTFT) and Time-Between-Tokens (TBT) SLOs. We present SuperInfer, a high-performance LLM inference system designed for emerging Superchips (e.g., NVIDIA GH200) with tightly coupled GPU-CPU architecture via NVLink-C2C. SuperInfer introduces RotaSched, the first proactive, SLO-aware rotary scheduler that rotates requests to maintain responsiveness on Superchips, and DuplexKV, an optimized rotation engine that enables full-duplex transfer over NVLink-C2C. Evaluations on GH200 using various models and datasets show that SuperInfer improves TTFT SLO attainment rates by up to 74.7% while maintaining comparable TBT and throughput compared to state-of-the-art systems, demonstrating that SLO-aware scheduling and memory co-design unlocks the full potential of Superchips for responsive LLM serving.
Authors: Zihan Zhou, Yinan Liu, Yuyang Xie, Bin Wang, Xiaochun Yang, Zezheng Feng
Abstract: The global shortage and uneven distribution of medical expertise continue to hinder equitable access to accurate diagnostic care. While existing intelligent diagnostic system have shown promise, most struggle with dual-user interaction, and dynamic knowledge integration -- limiting their real-world applicability. In this study, we present DiagLink, a dual-user diagnostic assistance system that synergizes large language models (LLMs), knowledge graphs (KGs), and medical experts to support both patients and physicians. DiagLink uses guided dialogues to elicit patient histories, leverages LLMs and KGs for collaborative reasoning, and incorporates physician oversight for continuous knowledge validation and evolution. The system provides a role-adaptive interface, dynamically visualized history, and unified multi-source evidence to improve both trust and usability. We evaluate DiagLink through user study, use cases and expert interviews, demonstrating its effectiveness in improving user satisfaction and diagnostic efficiency, while offering insights for the design of future AI-assisted diagnostic systems.
Authors: Zeyu Xing, Xing Li, Hui-Ling Zhen, Mingxuan Yuan, Sinno Jialin Pan
Abstract: KV caches, typically used only to speed up autoregressive decoding, encode contextual information that can be reused for downstream tasks at no extra cost. We propose treating the KV cache as a lightweight representation, eliminating the need to recompute or store full hidden states. Despite being weaker than dedicated embeddings, KV-derived representations are shown to be sufficient for two key applications: \textbf{(i) Chain-of-Embedding}, where they achieve competitive or superior performance on Llama-3.1-8B-Instruct and Qwen2-7B-Instruct; and \textbf{(ii) Fast/Slow Thinking Switching}, where they enable adaptive reasoning on Qwen3-8B and DeepSeek-R1-Distil-Qwen-14B, reducing token generation by up to $5.7\times$ with minimal accuracy loss. Our findings establish KV caches as a free, effective substrate for sampling and reasoning, opening new directions for representation reuse in LLM inference. Code: https://github.com/cmd2001/ICLR2026_KV-Embedding.
Authors: Brian Y. Tsui, Alan Y. Fang, Tiffany J. Hwu
Abstract: Robotic manipulation has increasingly adopted vision-language-action (VLA) models, which achieve strong performance but typically require task-specific demonstrations and fine-tuning, and often generalize poorly under domain shift. We investigate whether general-purpose large language model (LLM) agent frameworks, originally developed for software engineering, can serve as an alternative control paradigm for embodied manipulation. We introduce FAEA (Frontier Agent as Embodied Agent), which applies an LLM agent framework directly to embodied manipulation without modification. Using the same iterative reasoning that enables software agents to debug code, FAEA enables embodied agents to reason through manipulation strategies. We evaluate an unmodified frontier agent, Claude Agent SDK, across the LIBERO, ManiSkill3, and MetaWorld benchmarks. With privileged environment state access, FAEA achieves success rates of 84.9%, 85.7%, and 96%, respectively. This level of task success approaches that of VLA models trained with less than 100 demonstrations per task, without requiring demonstrations or fine-tuning. With one round of human feedback as an optional optimization, performance increases to 88.2% on LIBERO. This demonstration-free capability has immediate practical value: FAEA can autonomously explore novel scenarios in simulation and generate successful trajectories for training data augmentation in embodied learning. Our results indicate that general-purpose agents are sufficient for a class of manipulation tasks dominated by deliberative, task-level planning. This opens a path for robotics systems to leverage actively maintained agent infrastructure and benefit directly from ongoing advances in frontier models. Code is available at https://github.com/robiemusketeer/faea-sim
Authors: Qinzhuo Wu, Zhizhuo Yang, Hanhao Li, Pengzhi Gao, Wei Liu, Jian Luan
Abstract: Recent advances in mobile Graphical User Interface (GUI) agents highlight the growing need for comprehensive evaluation benchmarks. While new online benchmarks offer more realistic testing than offline ones, they tend to focus on the agents' task instruction-following ability while neglecting their reasoning and exploration ability. Moreover, these benchmarks do not consider the random noise in real-world mobile environments. This leads to a gap between benchmarks and real-world environments. To addressing these limitations, we propose MobileBench-OL, an online benchmark with 1080 tasks from 80 Chinese apps. It measures task execution, complex reasoning, and noise robustness of agents by including 5 subsets, which set multiple evaluation dimensions. We also provide an auto-eval framework with a reset mechanism, enabling stable and repeatable real-world benchmarking. Evaluating 12 leading GUI agents on MobileBench-OL shows significant room for improvement to meet real-world requirements. Human evaluation further confirms that MobileBench-OL can reliably measure the performance of leading GUI agents in real environments. Our data and code will be released upon acceptance.
Authors: Asifullah Khan, Aimen Wadood, Mubashar Iqbal, Umme Zahoora
Abstract: Ransomware has become one of the most serious cybersecurity threats causing major financial losses and operational disruptions worldwide.Traditional detection methods such as static analysis, heuristic scanning and behavioral analysis often fall short when used alone. To address these limitations, this paper presents multimodal multi agent ransomware analysis framework designed for ransomware classification. Proposed multimodal multiagent architecture combines information from static, dynamic and network sources. Each data type is handled by specialized agent that uses auto encoder based feature extraction. These representations are then integrated through a fusion agent. After that fused representation are used by transformer based classifier. It identifies the specific ransomware family. The agents interact through an interagent feedback mechanism that iteratively refines feature representations by suppressing low confidence information. The framework was evaluated on large scale datasets containing thousands of ransomware and benign samples. Multiple experiments were conducted on ransomware dataset. It outperforms single modality and nonadaptive fusion baseline achieving improvement of up to 0.936 in Macro-F1 for family classification and reducing calibration error. Over 100 epochs, the agentic feedback loop displays a stable monotonic convergence leading to over +0.75 absolute improvement in terms of agent quality and a final composite score of around 0.88 without fine tuning of the language models. Zeroday ransomware detection remains family dependent on polymorphism and modality disruptions. Confidence aware abstention enables reliable real world deployment by favoring conservativeand trustworthy decisions over forced classification. The findings indicate that proposed approach provides a practical andeffective path toward improving real world ransomware defense systems.
Authors: Yue Liang, Jiatong Du, Ziyi Yang, Yanjun Huang, Hong Chen
Abstract: Scene graphs provide structured abstractions for scene understanding, yet they often overfit to spurious correlations, severely hindering out-of-distribution generalization. To address this limitation, we propose CURVE, a causality-inspired framework that integrates variational uncertainty modeling with uncertainty-guided structural regularization to suppress high-variance, environment-specific relations. Specifically, we apply prototype-conditioned debiasing to disentangle invariant interaction dynamics from environment-dependent variations, promoting a sparse and domain-stable topology. Empirically, we evaluate CURVE in zero-shot transfer and low-data sim-to-real adaptation, verifying its ability to learn domain-stable sparse topologies and provide reliable uncertainty estimates to support risk prediction under distribution shifts.
Authors: Xiangbo Wang, Wenbin Jiang, Jin Wang, Yubo You, Sheng Fang, Fei Wen
Abstract: Recent neural audio compression models often rely on residual vector quantization for high-fidelity coding, but using a fixed number of per-frame codebooks is suboptimal for the wide variability of audio content-especially for signals that are either very simple or highly complex. To address this limitation, we propose SwitchCodec, a neural audio codec based on Residual Experts Vector Quantization (REVQ). REVQ combines a shared quantizer with dynamically routed expert quantizers that are activated according to the input audio, decoupling bitrate from codebook capacity and improving compression efficiency. This design ensures full training and utilization of each quantizer. In addition, a variable-bitrate mechanism adjusts the number of active expert quantizers at inference, enabling multi-bitrate operation without retraining. Experiments demonstrate that SwitchCodec surpasses existing baselines on both objective metrics and subjective listening tests.
Authors: Mariia Drozdova
Abstract: Can standard continuous-time generative models represent distributions whose support is an extremely sparse, globally constrained discrete set? We study this question using completed Sudoku grids as a controlled testbed, treating them as a subset of a continuous relaxation space. We train flow-matching and score-based models along a Gaussian probability path and compare deterministic (ODE) sampling, stochastic (SDE) sampling, and DDPM-style discretizations derived from the same continuous-time training. Unconditionally, stochastic sampling substantially outperforms deterministic flows; score-based samplers are the most reliable among continuous-time methods, and DDPM-style ancestral sampling achieves the highest validity overall. We further show that the same models can be repurposed for guided generation: by repeatedly sampling completions under clamped clues and stopping when constraints are satisfied, the model acts as a probabilistic Sudoku solver. Although far less sample-efficient than classical solvers and discrete-geometry-aware diffusion methods, these experiments demonstrate that classic diffusion/flow formulations can assign non-zero probability mass to globally constrained combinatorial structures and can be used for constraint satisfaction via stochastic search.
Authors: Wei Huang, Anda Cheng, Yinggui Wang, Lei Wang, Tao Wei
Abstract: Large Language Models (LLMs) can be fine-tuned on domain-specific data to enhance their performance in specialized fields. However, such data often contains numerous low-quality samples, necessitating effective data processing (DP). In practice, DP strategies are typically developed through iterative manual analysis and trial-and-error adjustment. These processes inevitably incur high labor costs and may lead to privacy issues in high-privacy domains like healthcare due to direct human access to sensitive data. Thus, achieving automated data processing without exposing the raw data has become a critical challenge. To address this challenge, we propose LLM-AutoDP, a novel framework that leverages LLMs as agents to automatically generate and optimize data processing strategies. Our method generates multiple candidate strategies and iteratively refines them using feedback signals and comparative evaluations. This iterative in-context learning mechanism enables the agent to converge toward high-quality processing pipelines without requiring direct human intervention or access to the underlying data. To further accelerate strategy search, we introduce three key techniques: Distribution Preserving Sampling, which reduces data volume while maintaining distributional integrity; Processing Target Selection, which uses a binary classifier to identify low-quality samples for focused processing; Cache-and-Reuse Mechanism}, which minimizes redundant computations by reusing prior processing results. Results show that models trained on data processed by our framework achieve over 80% win rates against models trained on unprocessed data. Compared to AutoML baselines based on LLM agents, LLM-AutoDP achieves approximately a 65% win rate. Moreover, our acceleration techniques reduce the total searching time by up to 10 times, demonstrating both effectiveness and efficiency.
Authors: Kaile Wang, Jiannong Cao, Yu Yang, Xiaoyin Li, Mingjin Zhang
Abstract: Heterogeneous federated learning (HFL) aims to ensure effective and privacy-preserving collaboration among different entities. As newly joined clients require significant adjustments and additional training to align with the existing system, the problem of generalizing federated learning models to unseen clients under heterogeneous data has become progressively crucial. Consequently, we highlight two unsolved challenging issues in federated domain generalization: Optimization Divergence and Performance Divergence. To tackle the above challenges, we propose FedRD, a novel heterogeneity-aware federated learning algorithm that collaboratively utilizes parameter-guided global generalization aggregation and local debiased classification to reduce divergences, aiming to obtain an optimal global model for participating and unseen clients. Extensive experiments on public multi-domain datasets demonstrate that our approach exhibits a substantial performance advantage over competing baselines in addressing this specific problem.
Authors: Ananya Shukla, Chaitanya Modi, Satvik Bajpai, Siddharth Siddharth
Abstract: Large Language Models (LLMs) have emerged as powerful learning tools, but they lack awareness of learners' cognitive and physiological states, limiting their adaptability to the user's learning style. Contemporary learning techniques primarily focus on structured learning paths, knowledge tracing, and generic adaptive testing but fail to address real-time learning challenges driven by cognitive load, attention fluctuations, and engagement levels. Building on findings from a formative user study (N=66), we introduce GuideAI, a multi-modal framework that enhances LLM-driven learning by integrating real-time biosensory feedback including eye gaze tracking, heart rate variability, posture detection, and digital note-taking behavior. GuideAI dynamically adapts learning content and pacing through cognitive optimizations (adjusting complexity based on learning progress markers), physiological interventions (breathing guidance and posture correction), and attention-aware strategies (redirecting focus using gaze analysis). Additionally, GuideAI supports diverse learning modalities, including text-based, image-based, audio-based, and video-based instruction, across varied knowledge domains. A preliminary study (N = 25) assessed GuideAI's impact on knowledge retention and cognitive load through standardized assessments. The results show statistically significant improvements in both problem-solving capability and recall-based knowledge assessments. Participants also experienced notable reductions in key NASA-TLX measures including mental demand, frustration levels, and effort, while simultaneously reporting enhanced perceived performance. These findings demonstrate GuideAI's potential to bridge the gap between current LLM-based learning systems and individualized learner needs, paving the way for adaptive, cognition-aware education at scale.
Authors: Jai Lal Lulla, Seyedmoein Mohsenimofidi, Matthias Galster, Jie M. Zhang, Sebastian Baltes, Christoph Treude
Abstract: AI coding agents such as Codex and Claude Code are increasingly used to autonomously contribute to software repositories. However, little is known about how repository-level configuration artifacts affect operational efficiency of the agents. In this paper, we study the impact of AGENTS.md files on the runtime and token consumption of AI coding agents operating on GitHub pull requests. We analyze 10 repositories and 124 pull requests, executing agents under two conditions: with and without an AGENTS.md file. We measure wall-clock execution time and token usage during agent execution. Our results show that the presence of AGENTS.md is associated with a lower median runtime ($\Delta 28.64$%) and reduced output token consumption ($\Delta 16.58$%), while maintaining a comparable task completion behavior. Based on these results, we discuss immediate implications for the configuration and deployment of AI coding agents in practice, and outline a broader research agenda on the role of repository-level instructions in shaping the behavior, efficiency, and integration of AI coding agents in software development workflows.
Authors: Nicholas Santavas, Kareem Eissa, Patrycja Cieplicka, Piotr Florek, Matteo Nulli, Stefan Vasilev, Seyyed Hadi Hashemi, Antonios Gasteratos, Shahram Khadivi
Abstract: Enterprise LLM deployment faces a critical scalability challenge: organizations must optimize models systematically to scale AI initiatives within constrained compute budgets, yet the specialized expertise required for manual optimization remains a niche and scarce skillset. This challenge is particularly evident in managing GPU utilization across heterogeneous infrastructure while enabling teams with diverse workloads and limited LLM optimization experience to deploy models efficiently. We present OptiKIT, a distributed LLM optimization framework that democratizes model compression and tuning by automating complex optimization workflows for non-expert teams. OptiKIT provides dynamic resource allocation, staged pipeline execution with automatic cleanup, and seamless enterprise integration. In production, it delivers more than 2x GPU throughput improvement while empowering application teams to achieve consistent performance improvements without deep LLM optimization expertise. We share both the platform design and key engineering insights into resource allocation algorithms, pipeline orchestration, and integration patterns that enable large-scale, production-grade democratization of model optimization. Finally, we open-source the system to enable external contributions and broader reproducibility.
Authors: Yuhao Sun, Chengyi Cai, Jiacheng Zhang, Zesheng Ye, Xingliang Yuan, Feng Liu
Abstract: Recent research has shown that aligning fine-grained text descriptions with localized image patches can significantly improve the zero-shot performance of pre-trained vision-language models (e.g., CLIP). However, we find that both fine-grained text descriptions and localized image patches often contain redundant information, making text-visual alignment less effective. In this paper, we tackle this issue from two perspectives: \emph{View Refinement} and \emph{Description refinement}, termed as \textit{\textbf{Bi}-refinement for \textbf{F}ine-grained \textbf{T}ext-visual \textbf{A}lignment} (BiFTA). \emph{View refinement} removes redundant image patches with high \emph{Intersection over Union} (IoU) ratios, resulting in more distinctive visual samples. \emph{Description refinement} removes redundant text descriptions with high pairwise cosine similarity, ensuring greater diversity in the remaining descriptions. BiFTA achieves superior zero-shot performance on 6 benchmark datasets for both ViT-based and ResNet-based CLIP, justifying the necessity to remove redundant information in visual-text alignment.
Authors: Yumou Liu, Zhenzhe Zheng, Jiang Rong, Yao Hu, Fan Wu, Guihai Chen
Abstract: Modern content platforms offer paid promotion to mitigate cold start by allocating exposure via auctions. Our empirical analysis reveals a counterintuitive flaw in this paradigm: while promotion rescues low-to-medium quality content, it can harm high-quality content by forcing exposure to suboptimal audiences, polluting engagement signals and downgrading future recommendation. We recast content promotion as a dual-objective optimization that balances short-term value acquisition with long-term model improvement. To make this tractable at bid time in content promotion, we introduce a decomposable surrogate objective, gradient coverage, and establish its formal connection to Fisher Information and optimal experimental design. We design a two-stage auto-bidding algorithm based on Lagrange duality that dynamically paces budget through a shadow price and optimizes impression-level bids using per-impression marginal utilities. To address missing labels at bid time, we propose a confidence-gated gradient heuristic, paired with a zeroth-order variant for black-box models that reliably estimates learning signals in real time. We provide theoretical guarantees, proving monotone submodularity of the composite objective, sublinear regret in online auction, and budget feasibility. Extensive offline experiments on synthetic and real-world datasets validate the framework: it outperforms baselines, achieves superior final AUC/LogLoss, adheres closely to budget targets, and remains effective when gradients are approximated zeroth-order. These results show that strategic, information-aware promotion can improve long-term model performance and organic outcomes beyond naive impression-maximization strategies.
Authors: Yigitcan \"Ozer, Wanying Ge, Zhe Zhang, Xin Wang, Junichi Yamagishi
Abstract: Audio watermarking embeds auxiliary information into speech while maintaining speaker identity, linguistic content, and perceptual quality. Although recent advances in neural and digital signal processing-based watermarking methods have improved imperceptibility and embedding capacity, robustness is still primarily assessed against conventional distortions such as compression, additive noise, and resampling. However, the rise of deep learning-based attacks introduces novel and significant threats to watermark security. In this work, we investigate self voice conversion as a universal, content-preserving attack against audio watermarking systems. Self voice conversion remaps a speaker's voice to the same identity while altering acoustic characteristics through a voice conversion model. We demonstrate that this attack severely degrades the reliability of state-of-the-art watermarking approaches and highlight its implications for the security of modern audio watermarking techniques.
Authors: Jiahe Li, Junru Chen, Fanqi Shen, Jialan Yang, Jada Li, Zhizhang Yuan, Baowen Cheng, Meng Li, Yang Yang
Abstract: Enabling natural communication through brain-computer interfaces (BCIs) remains one of the most profound challenges in neuroscience and neurotechnology. While existing frameworks offer partial solutions, they are constrained by oversimplified semantic representations and a lack of interpretability. To overcome these limitations, we introduce Semantic Intent Decoding (SID), a novel framework that translates neural activity into natural language by modeling meaning as a flexible set of compositional semantic units. SID is built on three core principles: semantic compositionality, continuity and expandability of semantic space, and fidelity in reconstruction. We present BrainMosaic, a deep learning architecture implementing SID. BrainMosaic decodes multiple semantic units from EEG/SEEG signals using set matching and then reconstructs coherent sentences through semantic-guided reconstruction. This approach moves beyond traditional pipelines that rely on fixed-class classification or unconstrained generation, enabling a more interpretable and expressive communication paradigm. Extensive experiments on multilingual EEG and clinical SEEG datasets demonstrate that SID and BrainMosaic offer substantial advantages over existing frameworks, paving the way for natural and effective BCI-mediated communication.
Authors: Fatima Ezzeddine, Obaida Ammar, Silvia Giordano, Omran Ayoub
Abstract: Explainable Artificial Intelligence (XAI) is becoming increasingly essential for enhancing the transparency of machine learning (ML) models. Among the various XAI techniques, counterfactual explanations (CFs) hold a pivotal role due to their ability to illustrate how changes in input features can alter an ML model's decision, thereby offering actionable recourse to users. Ensuring that individuals with comparable attributes and those belonging to different protected groups (e.g., demographic) receive similar and actionable recourse options is essential for trustworthy and fair decision-making. In this work, we address this challenge directly by focusing on the generation of fair CFs. Specifically, we start by defining and formulating fairness at: 1) individual fairness, ensuring that similar individuals receive similar CFs, 2) group fairness, ensuring equitable CFs across different protected groups and 3) hybrid fairness, which accounts for both individual and broader group-level fairness. We formulate the problem as an optimization task and propose a novel model-agnostic, reinforcement learning based approach to generate CFs that satisfy fairness constraints at both the individual and group levels, two objectives that are usually treated as orthogonal. As fairness metrics, we extend existing metrics commonly used for auditing ML models, such as equal choice of recourse and equal effectiveness across individuals and groups. We evaluate our approach on three benchmark datasets, showing that it effectively ensures individual and group fairness while preserving the quality of the generated CFs in terms of proximity and plausibility, and quantify the cost of fairness in the different levels separately. Our work opens a broader discussion on hybrid fairness and its role and implications for XAI and beyond CFs.
Authors: Jesse Phitidis, Alison Q. Smithard, William N. Whiteley, Joanna M. Wardlaw, Miguel O. Bernabeu, Maria Vald\'es Hern\'andez
Abstract: White matter hyperintensities (WMH) and ischaemic stroke lesions (ISL) are imaging features associated with cerebral small vessel disease (SVD) that are visible on brain magnetic resonance imaging (MRI) scans. The development and validation of deep learning models to segment and differentiate these features is difficult because they visually confound each other in the fluid-attenuated inversion recovery (FLAIR) sequence and often appear in the same subject. We investigated six strategies for training a combined WMH and ISL segmentation model using partially labelled data. We combined privately held fully and partially labelled datasets with publicly available partially labelled datasets to yield a total of 2052 MRI volumes, with 1341 and 1152 containing ground truth annotations for WMH and ISL respectively. We found that several methods were able to effectively leverage the partially labelled data to improve model performance, with the use of pseudolabels yielding the best result.
Authors: Robin Singh, Aditya Yogesh Nair, Fabio Palumbo, Florian Barbaro, Anna Dyka, Lohith Rachakonda
Abstract: Recent advances in Text-to-Speech (TTS) systems have substantially increased the realism of synthetic speech, raising new challenges for audio deepfake detection. This work presents a comparative evaluation of three state-of-the-art TTS models--Dia2, Maya1, and MeloTTS--representing streaming, LLM-based, and non-autoregressive architectures. A corpus of 12,000 synthetic audio samples was generated using the Daily-Dialog dataset and evaluated against four detection frameworks, including semantic, structural, and signal-level approaches. The results reveal significant variability in detector performance across generative mechanisms: models effective against one TTS architecture may fail against others, particularly LLM-based synthesis. In contrast, a multi-view detection approach combining complementary analysis levels demonstrates robust performance across all evaluated models. These findings highlight the limitations of single-paradigm detectors and emphasize the necessity of integrated detection strategies to address the evolving landscape of audio deepfake threats.
Authors: Jiawen Chen, Qi Shao, Mingtong Zhou, Duxin Chen, Wenwu Yu
Abstract: Topological deep learning has emerged for modeling higher-order relational structures beyond pairwise interactions that standard graph neural networks fail to capture. Although combinatorial complexes offer a unified topological framework, most existing topological deep learning methods rely on local message passing via attention mechanisms, which incur quadratic complexity and remain low-dimensional, limiting scalability and rank-aware information aggregation in higher-order complexes.We propose Combinatorial Complex Mamba (CCMamba), the first unified mamba-based neural framework for learning on combinatorial complexes. CCMamba reformulates message passing as a selective state-space modeling problem by organizing multi-rank incidence relations into structured sequences processed by rank-aware state-space models. This enables adaptive, directional, and long range information propagation in linear time without self attention. We further establish the theoretical analysis that the expressive power upper-bound of CCMamba message passing is the 1-Weisfeiler-Lehman test. Experiments on graph, hypergraph, and simplicial benchmarks demonstrate that CCMamba consistently outperforms existing methods while exhibiting improved scalability and robustness to depth.
Authors: Ling Tang, Jilin Mei, Dongrui Liu, Chen Qian, Dawei Cheng, Jing Shao, Xia Hu
Abstract: Large language model-powered multi-agent systems have emerged as powerful tools for simulating complex human-like systems. The interactions within these systems often lead to extreme events whose origins remain obscured by the black box of emergence. Interpreting these events is critical for system safety. This paper proposes the first framework for explaining emergent extreme events in multi-agent systems, aiming to answer three fundamental questions: When does the event originate? Who drives it? And what behaviors contribute to it? Specifically, we adapt the Shapley value to faithfully attribute the occurrence of extreme events to each action taken by agents at different time steps, i.e., assigning an attribution score to the action to measure its influence on the event. We then aggregate the attribution scores along the dimensions of time, agent, and behavior to quantify the risk contribution of each dimension. Finally, we design a set of metrics based on these contribution scores to characterize the features of extreme events. Experiments across diverse multi-agent system scenarios (economic, financial, and social) demonstrate the effectiveness of our framework and provide general insights into the emergence of extreme phenomena.
Authors: Kahraman Kostas, Rabia Yasa Kostas
Abstract: This paper critically examines the device identification process using machine learning, addressing common pitfalls in existing literature. We analyze the trade-offs between identification methods (unique vs. class based), data heterogeneity, feature extraction challenges, and evaluation metrics. By highlighting specific errors, such as improper data augmentation and misleading session identifiers, we provide a robust guideline for researchers to enhance the reproducibility and generalizability of IoT security models.
Authors: Ariel Maymon, Yanir Buznah, Uri Shaham
Abstract: Unsupervised ensemble learning emerged to address the challenge of combining multiple learners' predictions without access to ground truth labels or additional data. This paradigm is crucial in scenarios where evaluating individual classifier performance or understanding their strengths is challenging due to limited information. We propose a novel deep energy-based method for constructing an accurate meta-learner using only the predictions of individual learners, potentially capable of capturing complex dependence structures between them. Our approach requires no labeled data, learner features, or problem-specific information, and has theoretical guarantees for when learners are conditionally independent. We demonstrate superior performance across diverse ensemble scenarios, including challenging mixture of experts settings. Our experiments span standard ensemble datasets and curated datasets designed to test how the model fuses expertise from multiple sources. These results highlight the potential of unsupervised ensemble learning to harness collective intelligence, especially in data-scarce or privacy-sensitive environments.
Authors: Anna van Elst, Igor Colin, Stephan Cl\'emen\c{c}on
Abstract: Specifications for decentralized learning on resource-constrained edge devices require algorithms that are communication-efficient, robust to data corruption, and lightweight in memory usage. While state-of-the-art gossip-based methods satisfy the first requirement, achieving robustness remains challenging. Asynchronous decentralized ADMM-based methods have been explored for estimating the median, a statistical centrality measure that is notoriously more robust than the mean. However, existing approaches require memory that scales with node degree, making them impractical when memory is limited. In this paper, we propose AsylADMM, a novel gossip algorithm for decentralized median and quantile estimation, primarily designed for asynchronous updates and requiring only two variables per node. We analyze a synchronous variant of AsylADMM to establish theoretical guarantees and empirically demonstrate fast convergence for the asynchronous algorithm. We then show that our algorithm enables quantile-based trimming, geometric median estimation, and depth-based trimming, with quantile-based trimming empirically outperforming existing rank-based methods. Finally, we provide a novel theoretical analysis of rank-based trimming via Markov chain theory.
Authors: Dimitris Michailidis, Sennay Ghebreab, Fernando P. Santos
Abstract: Who benefits from expanding transport networks? While designed to improve mobility, such interventions can also create inequality. In this paper, we show that disparities arise not only from the structure of the network itself but also from differences in how commuters adapt to it. We model commuters as reinforcement learning agents who adapt their travel choices at different learning rates, reflecting unequal access to resources and information. To capture potential efficiency-fairness tradeoffs, we introduce the Price of Learning (PoL), a measure of inefficiency during learning. We analyze both a stylized network -- inspired in the well-known Braess's paradox, yet with two-source nodes -- and an abstraction of a real-world metro system (Amsterdam). Our simulations show that network expansions can simultaneously increase efficiency and amplify inequality, especially when faster learners disproportionately benefit from new routes before others adapt. These results highlight that transport policies must account not only for equilibrium outcomes but also for the heterogeneous ways commuters adapt, since both shape the balance between efficiency and fairness.
Authors: Aiming Hao, Chen Zhu, Jiashu Zhu, Jiahong Wu, Xiangxiang Chu
Abstract: Ordinal regression and ranking are challenging due to inherent ordinal dependencies that conventional methods struggle to model. We propose Ranking-Aware Reinforcement Learning (RARL), a novel RL framework that explicitly learns these relationships. At its core, RARL features a unified objective that synergistically integrates regression and Learning-to-Rank (L2R), enabling mutual improvement between the two tasks. This is driven by a ranking-aware verifiable reward that jointly assesses regression precision and ranking accuracy, facilitating direct model updates via policy optimization. To further enhance training, we introduce Response Mutation Operations (RMO), which inject controlled noise to improve exploration and prevent stagnation at saddle points. The effectiveness of RARL is validated through extensive experiments on three distinct benchmarks.
Authors: Lakshman Balasubramanian
Abstract: Person Re-Identification (ReID) remains a challenging problem in computer vision. This work reviews various training paradigm and evaluates the robustness of state-of-the-art ReID models in cross-domain applications and examines the role of foundation models in improving generalization through richer, more transferable visual representations. We compare three training paradigms, supervised, self-supervised, and language-aligned models. Through the study the aim is to answer the following questions: Can supervised models generalize in cross-domain scenarios? How does foundation models like SigLIP2 perform for the ReID tasks? What are the weaknesses of current supervised and foundational models for ReID? We have conducted the analysis across 11 models and 9 datasets. Our results show a clear split: supervised models dominate their training domain but crumble on cross-domain data. Language-aligned models, however, show surprising robustness cross-domain for ReID tasks, even though they are not explicitly trained to do so. Code and data available at: https://github.com/moiiai-tech/object-reid-benchmark.
Authors: Hyunjun Na, Donghwan Lee
Abstract: Gradient temporal-difference (GTD) learning algorithms are widely used for off-policy policy evaluation with function approximation. However, existing convergence analyses rely on the restrictive assumption that the so-called feature interaction matrix (FIM) is nonsingular. In practice, the FIM can become singular and leads to instability or degraded performance. In this paper, we propose a regularized optimization objective by reformulating the mean-square projected Bellman error (MSPBE) minimization. This formulation naturally yields a regularized GTD algorithms, referred to as R-GTD, which guarantees convergence to a unique solution even when the FIM is singular. We establish theoretical convergence guarantees and explicit error bounds for the proposed method, and validate its effectiveness through empirical experiments.
Authors: Zhuonan Wang, Wenjie Yan, Wenqiao Zhang, Xiaohui Song, Jian Ma, Ke Yao, Yibo Yu, Beng Chin Ooi
Abstract: Medical image classification is a core task in computer-aided diagnosis (CAD), playing a pivotal role in early disease detection, treatment planning, and patient prognosis assessment. In ophthalmic practice, fluorescein fundus angiography (FFA) and indocyanine green angiography (ICGA) provide hemodynamic and lesion-structural information that conventional fundus photography cannot capture. However, due to the single-modality nature, subtle lesion patterns, and significant inter-device variability, existing methods still face limitations in generalization and high-confidence prediction. To address these challenges, we propose CLEAR-Mamba, an enhanced framework built upon MedMamba with optimizations in both architecture and training strategy. Architecturally, we introduce HaC, a hypernetwork-based adaptive conditioning layer that dynamically generates parameters according to input feature distributions, thereby improving cross-domain adaptability. From a training perspective, we develop RaP, a reliability-aware prediction scheme built upon evidential uncertainty learning, which encourages the model to emphasize low-confidence samples and improves overall stability and reliability. We further construct a large-scale ophthalmic angiography dataset covering both FFA and ICGA modalities, comprising multiple retinal disease categories for model training and evaluation. Experimental results demonstrate that CLEAR-Mamba consistently outperforms multiple baseline models, including the original MedMamba, across various metrics-showing particular advantages in multi-disease classification and reliability-aware prediction. This study provides an effective solution that balances generalizability and reliability for modality-specific medical image classification tasks.
Authors: Xinyu Wang, Ruoyu Wang, Qiangwei Peng, Peijie Zhou, Tiejun Li
Abstract: Reconstructing dynamical evolution from limited observations is a fundamental challenge in single-cell biology, where dynamic unbalanced optimal transport provides a principled framework for modeling coupled transport and mass variation. However, existing approaches rely on trajectory simulation at inference time, making inference a key bottleneck for scalable applications. In this work, we propose a mean-flow framework for unbalanced flow matching that summarizes both transport and mass-growth dynamics over arbitrary time intervals using mean velocity and mass-growth fields, enabling fast one-step generation without trajectory simulation. To solve dynamic unbalanced optimal transport under the Wasserstein-Fisher-Rao geometry, we further build on this framework to develop Wasserstein-Fisher-Rao Mean Flow Matching (WFR-MFM). Across synthetic and real single-cell RNA sequencing datasets, WFR-MFM achieves orders-of-magnitude faster inference than a range of existing baselines while maintaining high predictive accuracy, and enables efficient perturbation response prediction on large synthetic datasets with thousands of conditions.
Authors: Jonathan Rystr{\o}m, Chris Schmitz, Karolina Korgul, Jan Batzner, Chris Russell
Abstract: Deploying Large Language Model-based agents (LLM agents) in the public sector requires assuring that they meet the stringent legal, procedural, and structural requirements of public-sector institutions. Practitioners and researchers often turn to benchmarks for such assessments. However, it remains unclear what criteria benchmarks must meet to ensure they adequately reflect public-sector requirements, or how many existing benchmarks do so. In this paper, we first define such criteria based on a first-principles survey of public administration literature: benchmarks must be \emph{process-based}, \emph{realistic}, \emph{public-sector-specific} and report \emph{metrics} that reflect the unique requirements of the public sector. We analyse more than 1,300 benchmark papers for these criteria using an expert-validated LLM-assisted pipeline. Our results show that no single benchmark meets all of the criteria. Our findings provide a call to action for both researchers to develop public sector-relevant benchmarks and for public-sector officials to apply these criteria when evaluating their own agentic use cases.
Authors: Shuguang Zhang, Junhong Lian, Guoxin Yu, Baoxun Xu, Xiang Ao
Abstract: Multimodal sarcasm detection (MSD) aims to identify sarcasm within image-text pairs by modeling semantic incongruities across modalities. Existing methods often exploit cross-modal embedding misalignment to detect inconsistency but struggle when visual and textual content are loosely related or semantically indirect. While recent approaches leverage large language models (LLMs) to generate sarcastic cues, the inherent diversity and subjectivity of these generations often introduce noise. To address these limitations, we propose the Generative Discrepancy Comparison Network (GDCNet). This framework captures cross-modal conflicts by utilizing descriptive, factually grounded image captions generated by Multimodal LLMs (MLLMs) as stable semantic anchors. Specifically, GDCNet computes semantic and sentiment discrepancies between the generated objective description and the original text, alongside measuring visual-textual fidelity. These discrepancy features are then fused with visual and textual representations via a gated module to adaptively balance modality contributions. Extensive experiments on MSD benchmarks demonstrate GDCNet's superior accuracy and robustness, establishing a new state-of-the-art on the MMSD2.0 benchmark.
Authors: Rohan Asthana, Vasileios Belagiannis
Abstract: Diffusion-based image generative models produce high-fidelity images through iterative denoising but remain vulnerable to memorization, where they unintentionally reproduce exact copies or parts of training images. Recent memorization detection methods are primarily based on the norm of score difference as indicators of memorization. We prove that such norm-based metrics are mainly effective under the assumption of isotropic log-probability distributions, which generally holds at high or medium noise levels. In contrast, analyzing the anisotropic regime reveals that memorized samples exhibit strong angular alignment between the guidance vector and unconditional scores in the low-noise setting. Through these insights, we develop a memorization detection metric by integrating isotropic norm and anisotropic alignment. Our detection metric can be computed directly on pure noise inputs via two conditional and unconditional forward passes, eliminating the need for costly denoising steps. Detection experiments on Stable Diffusion v1.4 and v2 show that our metric outperforms existing denoising-free detection methods while being at least approximately 5x faster than the previous best approach. Finally, we demonstrate the effectiveness of our approach by utilizing a mitigation strategy that adapts memorized prompts based on our developed metric.
Authors: Alejandro Luque-Cerpa, Mengyuan Wang, Emil Carlsson, Sanjit A. Seshia, Devdatt Dubhashi, Hazem Torfah
Abstract: We introduce a novel framework for learning context-aware runtime monitors for AI-based control ensembles. Machine-learning (ML) controllers are increasingly deployed in (autonomous) cyber-physical systems because of their ability to solve complex decision-making tasks. However, their accuracy can degrade sharply in unfamiliar environments, creating significant safety concerns. Traditional ensemble methods aim to improve robustness by averaging or voting across multiple controllers, yet this often dilutes the specialized strengths that individual controllers exhibit in different operating contexts. We argue that, rather than blending controller outputs, a monitoring framework should identify and exploit these contextual strengths. In this paper, we reformulate the design of safe AI-based control ensembles as a contextual monitoring problem. A monitor continuously observes the system's context and selects the controller best suited to the current conditions. To achieve this, we cast monitor learning as a contextual learning task and draw on techniques from contextual multi-armed bandits. Our approach comes with two key benefits: (1) theoretical safety guarantees during controller selection, and (2) improved utilization of controller diversity. We validate our framework in two simulated autonomous driving scenarios, demonstrating significant improvements in both safety and performance compared to non-contextual baselines.
Authors: Juan Jose Rubio Jan, Jack Wu, Julia Ive
Abstract: This study applies Large Language Models (LLMs) to two foundational Electronic Health Record (EHR) data science tasks: structured data querying (using programmatic languages, Python/Pandas) and information extraction from unstructured clinical text via a Retrieval Augmented Generation (RAG) pipeline. We test the ability of LLMs to interact accurately with large structured datasets for analytics and the reliability of LLMs in extracting semantically correct information from free text health records when supported by RAG. To this end, we presented a flexible evaluation framework that automatically generates synthetic question and answer pairs tailored to the characteristics of each dataset or task. Experiments were conducted on a curated subset of MIMIC III, (four structured tables and one clinical note type), using a mix of locally hosted and API-based LLMs. Evaluation combined exact-match metrics, semantic similarity, and human judgment. Our findings demonstrate the potential of LLMs to support precise querying and accurate information extraction in clinical workflows.
Authors: Xinyue Li, Zhichao Zhang, Zhiming Xu, Shubo Xu, Xiongkuo Min, Yitong Chen, Guangtao Zhai
Abstract: Recent multimodal large language models (MLLMs) have demonstrated strong capabilities in image quality assessment (IQA) tasks. However, adapting such large-scale models is computationally expensive and still relies on substantial Mean Opinion Score (MOS) annotations. We argue that for MLLM-based IQA, the core bottleneck lies not in the quality perception capacity of MLLMs, but in MOS scale calibration. Therefore, we propose LEAF, a Label-Efficient Image Quality Assessment Framework that distills perceptual quality priors from an MLLM teacher into a lightweight student regressor, enabling MOS calibration with minimal human supervision. Specifically, the teacher conducts dense supervision through point-wise judgments and pair-wise preferences, with an estimate of decision reliability. Guided by these signals, the student learns the teacher's quality perception patterns through joint distillation and is calibrated on a small MOS subset to align with human annotations. Experiments on both user-generated and AI-generated IQA benchmarks demonstrate that our method significantly reduces the need for human annotations while maintaining strong MOS-aligned correlations, making lightweight IQA practical under limited annotation budgets.
Authors: Zhuang Yu, Lei Shen, Jing Zhao, Shiliang Sun
Abstract: Recent multimodal large language models (MLLMs) have shown remarkable progress across vision, audio, and language tasks, yet their performance on long-form, knowledge-intensive, and temporally structured educational content remains largely unexplored. To bridge this gap, we introduce LEMON, a Lecture-based Evaluation benchmark for MultimOdal uNderstanding, focusing on STEM lecture videos that require long-horizon reasoning and cross-modal integration. LEMON comprises 2,277 video segments spanning 5 disciplines and 29 courses, with an average duration of 196.1 seconds, yielding 4,181 high-quality QA pairs, including 3,413 multiple-choice and 768 open-ended questions. Distinct from existing video benchmarks, LEMON features: (1) semantic richness and disciplinary density, (2) tightly coupled video-audio-text modalities, (3) explicit temporal and pedagogical structure, and (4) contextually linked multi-turn questioning. It further encompasses six major tasks and twelve subtasks, covering the full cognitive spectrum from perception to reasoning and then to generation. Comprehensive experiments reveal substantial performance gaps across tasks, highlighting that even state-of-the-art MLLMs like GPT-4o struggle with temporal reasoning and instructional prediction. We expect LEMON to serve as an extensible and challenging benchmark for advancing multimodal perception, reasoning, and generation in long-form instructional contents.
Authors: Binglei Lou, Haoran Wu, Yao Lai, Jiayi Nie, Can Xiao, Xuan Guo, Rika Antonova, Robert Mullins, Aaron Zhao
Abstract: Diffusion Large Language Models (dLLMs) introduce iterative denoising to enable parallel token generation, but their sampling phase displays fundamentally different characteristics compared to GEMM-centric transformer layers. Profiling on modern GPUs reveals that sampling can account for up to 70% of total model inference latency-primarily due to substantial memory loads and writes from vocabulary-wide logits, reduction-based token selection, and iterative masked updates. These processes demand large on-chip SRAM and involve irregular memory accesses that conventional NPUs struggle to handle efficiently. To address this, we identify a set of critical instructions that an NPU architecture must specifically optimize for dLLM sampling. Our design employs lightweight non-GEMM vector primitives, in-place memory reuse strategies, and a decoupled mixed-precision memory hierarchy. Together, these optimizations deliver up to a 2.53x speedup over the NVIDIA RTX A6000 GPU under an equivalent nm technology node. We also open-source our cycle-accurate simulation and post-synthesis RTL verification code, confirming functional equivalence with current dLLM PyTorch implementations.
Authors: Raul de la Rosa, Ivana Dusparic, Nicolas Cardozo
Abstract: Reinforcement Learning (RL) agents often struggle in real-world applications where environmental conditions are non-stationary, particularly when reward functions shift or the available action space expands. This paper introduces MORPHIN, a self-adaptive Q-learning framework that enables on-the-fly adaptation without full retraining. By integrating concept drift detection with dynamic adjustments to learning and exploration hyperparameters, MORPHIN adapts agents to changes in both the reward function and on-the-fly expansions of the agent's action space, while preserving prior policy knowledge to prevent catastrophic forgetting. We validate our approach using a Gridworld benchmark and a traffic signal control simulation. The results demonstrate that MORPHIN achieves superior convergence speed and continuous adaptation compared to a standard Q-learning baseline, improving learning efficiency by up to 1.7x.
Authors: Matej Halinkovic, Nina Masarykova, Alexey Vinel, Marek Galinski
Abstract: End-to-end perception and trajectory prediction from raw sensor data is one of the key capabilities for autonomous driving. Modular pipelines restrict information flow and can amplify upstream errors. Recent query-based, fully differentiable perception-and-prediction (PnP) models mitigate these issues, yet the complementarity of cameras and LiDAR in the query-space has not been sufficiently explored. Models often rely on fusion schemes that introduce heuristic alignment and discrete selection steps which prevent full utilization of available information and can introduce unwanted bias. We propose Li-ViP3D++, a query-based multimodal PnP framework that introduces Query-Gated Deformable Fusion (QGDF) to integrate multi-view RGB and LiDAR in query space. QGDF (i) aggregates image evidence via masked attention across cameras and feature levels, (ii) extracts LiDAR context through fully differentiable BEV sampling with learned per-query offsets, and (iii) applies query-conditioned gating to adaptively weight visual and geometric cues per agent. The resulting architecture jointly optimizes detection, tracking, and multi-hypothesis trajectory forecasting in a single end-to-end model. On nuScenes, Li-ViP3D++ improves end-to-end behavior and detection quality, achieving higher EPA (0.335) and mAP (0.502) while substantially reducing false positives (FP ratio 0.147), and it is faster than the prior Li-ViP3D variant (139.82 ms vs. 145.91 ms). These results indicate that query-space, fully differentiable camera-LiDAR fusion can increase robustness of end-to-end PnP without sacrificing deployability.
Authors: Mae Sosto, Delfina Sol Martinez Pandiani, Laura Hollink
Abstract: This paper examines how Large Language Models (LLMs) reproduce societal norms, particularly heterocisnormativity, and how these norms translate into measurable biases in their text generations. We investigate whether explicit information about a subject's gender or sexuality influences LLM responses across three subject categories: queer-marked, non-queer-marked, and the normalized "unmarked" category. Representational imbalances are operationalized as measurable differences in English sentence completions across four dimensions: sentiment, regard, toxicity, and prediction diversity. Our findings show that Masked Language Models (MLMs) produce the least favorable sentiment, higher toxicity, and more negative regard for queer-marked subjects. Autoregressive Language Models (ARLMs) partially mitigate these patterns, while closed-access ARLMs tend to produce more harmful outputs for unmarked subjects. Results suggest that LLMs reproduce normative social assumptions, though the form and degree of bias depend strongly on specific model characteristics, which may redistribute, but not eliminate, representational harms.
Authors: Guoan Wang, Feiyu Wang, Zongwei Lv, Yikun Zong, Tong Yang
Abstract: As large language models (LLMs) continue to scale, deployment is increasingly bottlenecked by the memory wall, motivating a shift toward extremely low-bit quantization. However, most quantization-aware training (QAT) methods apply hard rounding and the straight-through estimator (STE) from the beginning of the training, which prematurely discretizes the optimization landscape and induces persistent gradient mismatch between latent weights and quantized weights, hindering effective optimization of quantized models. To address this, we propose Hestia, a Hessian-guided differentiable QAT framework for extremely low-bit LLMs, which replaces the rigid step function with a temperature-controlled softmax relaxation to maintain gradient flow early in training while progressively hardening quantization. Furthermore, Hestia leverages a tensor-wise Hessian trace metric as a lightweight curvature signal to drive fine-grained temperature annealing, enabling sensitivity-aware discretization across the model. Evaluations on Llama-3.2 show that Hestia consistently outperforms existing ternary QAT baselines, yielding average zero-shot improvements of 5.39% and 4.34% for the 1B and 3B models. These results indicate that Hessian-guided relaxation effectively recovers representational capacity, establishing a more robust training path for 1.58-bit LLMs. The code is available at https://github.com/hestia2026/Hestia.
Authors: Th\'eo Delemazure
Abstract: In an ordinal election, two candidates are said to be perfect clones if every voter ranks them adjacently. The independence of clones axiom then states that removing one of the two clones should not change the election outcome. This axiom has been extensively studied in social choice theory, and several voting rules are known to satisfy it (such as IRV, Ranked Pairs and Schulze). However, perfect clones are unlikely to occur in practice, especially for political elections with many voters. In this work, we study different notions of approximate clones in ordinal elections. Informally, two candidates are approximate clones in a preference profile if they are close to being perfect clones. We discuss two measures to quantify this proximity, and we show under which conditions the voting rules that are known to be independent of clones are also independent of approximate clones. In particular, we show that for elections with at least four candidates, none of these rules are independent of approximate clones in the general case. However, we find a more positive result for the case of three candidates. Finally, we conduct an empirical study of approximate clones and independence of approximate clones based on three real-world datasets: votes in local Scottish elections, votes in mini-jury deliberations, and votes of judges in figure skating competitions. We find that approximate clones are common in some contexts, and that the closest two candidates are to being perfect clones, the less likely their removal is to change the election outcome, especially for voting rules that are independent of perfect clones.
Authors: Haonan Zhong, Wei Song, Tingxu Han, Maurice Pagnucco, Jingling Xue, Yang Song
Abstract: Text-to-video (T2V) diffusion models have achieved rapid progress, yet their demographic biases, particularly gender bias, remain largely unexplored. We present FairT2V, a training-free debiasing framework for text-to-video generation that mitigates encoder-induced bias without finetuning. We first analyze demographic bias in T2V models and show that it primarily originates from pretrained text encoders, which encode implicit gender associations even for neutral prompts. We quantify this effect with a gender-leaning score that correlates with bias in generated videos. Based on this insight, FairT2V mitigates demographic bias by neutralizing prompt embeddings via anchor-based spherical geodesic transformations while preserving semantics. To maintain temporal coherence, we apply debiasing only during early identity-forming steps through a dynamic denoising schedule. We further propose a video-level fairness evaluation protocol combining VideoLLM-based reasoning with human verification. Experiments on the modern T2V model Open-Sora show that FairT2V substantially reduces demographic bias across occupations with minimal impact on video quality.
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 importance estimates in conditional settings, whereas condPED-ANOVA consistently provides meaningful importances that reflect the underlying conditional structure.
Authors: Jonas H\"ubotter, Frederike L\"ubeck, Lejs Behric, Anton Baumann, Marco Bagatella, Daniel Marta, Ido Hakimi, Idan Shenfeld, Thomas Kleine Buening, Carlos Guestrin, Andreas Krause
Abstract: Large language models are increasingly post-trained with reinforcement learning in verifiable domains such as code and math. Yet, current methods for reinforcement learning with verifiable rewards (RLVR) learn only from a scalar outcome reward per attempt, creating a severe credit-assignment bottleneck. Many verifiable environments actually provide rich textual feedback, such as runtime errors or judge evaluations, that explain why an attempt failed. We formalize this setting as reinforcement learning with rich feedback and introduce Self-Distillation Policy Optimization (SDPO), which converts tokenized feedback into a dense learning signal without any external teacher or explicit reward model. SDPO treats the current model conditioned on feedback as a self-teacher and distills its feedback-informed next-token predictions back into the policy. In this way, SDPO leverages the model's ability to retrospectively identify its own mistakes in-context. Across scientific reasoning, tool use, and competitive programming on LiveCodeBench v6, SDPO improves sample efficiency and final accuracy over strong RLVR baselines. Notably, SDPO also outperforms baselines in standard RLVR environments that only return scalar feedback by using successful rollouts as implicit feedback for failed attempts. Finally, applying SDPO to individual questions at test time accelerates discovery on difficult binary-reward tasks, achieving the same discovery probability as best-of-k sampling or multi-turn conversations with 3x fewer attempts.
Authors: Steve Azzolin, Stefano Teso, Bruno Lepri, Andrea Passerini, Sagar Malhotra
Abstract: Explanations provided by Self-explainable Graph Neural Networks (SE-GNNs) are fundamental for understanding the model's inner workings and for identifying potential misuse of sensitive attributes. Although recent works have highlighted that these explanations can be suboptimal and potentially misleading, a characterization of their failure cases is unavailable. In this work, we identify a critical failure of SE-GNN explanations: explanations can be unambiguously unrelated to how the SE-GNNs infer labels. We show that, on the one hand, many SE-GNNs can achieve optimal true risk while producing these degenerate explanations, and on the other, most faithfulness metrics can fail to identify these failure modes. Our empirical analysis reveals that degenerate explanations can be maliciously planted (allowing an attacker to hide the use of sensitive attributes) and can also emerge naturally, highlighting the need for reliable auditing. To address this, we introduce a novel faithfulness metric that reliably marks degenerate explanations as unfaithful, in both malicious and natural settings. Our code is available in the supplemental.
Authors: Minwu Kim, Safal Shrestha, Keith Ross
Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has substantially improved the reasoning abilities of large language models (LLMs), yet training often stalls as problems become saturated. We identify the core challenge as the poor accessibility of informative failures: learning signals exist but are rarely encountered during standard rollouts. To address this, we propose failure-prefix conditioning, a simple and effective method for learning from saturated problems. Rather than starting from the original question, our approach reallocates exploration by conditioning training on prefixes derived from rare incorrect reasoning trajectories, thereby exposing the model to failure-prone states. We observe that failure-prefix conditioning yields performance gains matching those of training on medium-difficulty problems, while preserving token efficiency. Furthermore, we analyze the model's robustness, finding that our method reduces performance degradation under misleading failure prefixes, albeit with a mild trade-off in adherence to correct early reasoning. Finally, we demonstrate that an iterative approach, which refreshes failure prefixes during training, unlocks additional gains after performance plateaus. Overall, our results suggest that failure-prefix conditioning offers an effective pathway to extend RLVR training on saturated problems.
Authors: Jie Liu, Yu Sun, Alpar Cseke, Yao Feng, Nicolas Heron, Michael J. Black, Yan Zhang
Abstract: Generating 3D humans that functionally interact with 3D scenes remains an open problem with applications in embodied AI, robotics, and interactive content creation. The key challenge involves reasoning about both the semantics of functional elements in 3D scenes and the 3D human poses required to achieve functionality-aware interaction. Unfortunately, existing methods typically lack explicit reasoning over object functionality and the corresponding human-scene contact, resulting in implausible or functionally incorrect interactions. In this work, we propose FunHSI, a training-free, functionality-driven framework that enables functionally correct human-scene interactions from open-vocabulary task prompts. Given a task prompt, FunHSI performs functionality-aware contact reasoning to identify functional scene elements, reconstruct their 3D geometry, and model high-level interactions via a contact graph. We then leverage vision-language models to synthesize a human performing the task in the image and estimate proposed 3D body and hand poses. Finally, the proposed 3D body configuration is refined via stage-wise optimization to ensure physical plausibility and functional correctness. In contrast to existing methods, FunHSI not only synthesizes more plausible general 3D interactions, such as "sitting on a sofa'', while supporting fine-grained functional human-scene interactions, e.g., "increasing the room temperature''. Extensive experiments demonstrate that FunHSI consistently generates functionally correct and physically plausible human-scene interactions across diverse indoor and outdoor scenes.
Authors: Brian Christian, Jessica A. F. Thompson, Elle Michelle Yang, Vincent Adam, Hannah Rose Kirk, Christopher Summerfield, Tsvetomira Dumbalska
Abstract: Reward models (RMs) are central to aligning large language models (LLMs) with human values but have received less attention than pre-trained and post-trained LLMs themselves. Because RMs are initialized from LLMs, they inherit representations that shape their behavior, but the nature and extent of this influence remain understudied. In a comprehensive study of 10 leading open-weight RMs using validated psycholinguistic corpora, we show that RMs exhibit significant differences along multiple dimensions of human value as a function of their base model. Using the "Big Two" psychological axes, we show a robust preference of Llama RMs for "agency" and a corresponding robust preference of Gemma RMs for "communion." This phenomenon holds even when the preference data and finetuning process are identical, and we trace it back to the logits of the respective instruction-tuned and pre-trained models. These log-probability differences themselves can be formulated as an implicit RM; we derive usable implicit reward scores and show that they exhibit the very same agency/communion difference. We run experiments training RMs with ablations for preference data source and quantity, which demonstrate that this effect is not only repeatable but surprisingly durable. Despite RMs being designed to represent human preferences, our evidence shows that their outputs are influenced by the pretrained LLMs on which they are based. This work underscores the importance of safety and alignment efforts at the pretraining stage, and makes clear that open-source developers' choice of base model is as much a consideration of values as of performance.
Authors: Zihao Wang, Hang Yin, Lihui Liu, Hanghang Tong, Yangqiu Song, Ginny Wong, Simon See
Abstract: This paper studies the minimal dimension required to embed subset memberships ($m$ elements and ${m\choose k}$ subsets of at most $k$ elements) into vector spaces, denoted as Minimal Embeddable Dimension (MED). The tight bounds of MED are derived theoretically and supported empirically for various notions of "distances" or "similarities," including the $\ell_2$ metric, inner product, and cosine similarity. In addition, we conduct numerical simulation in a more achievable setting, where the ${m\choose k}$ subset embeddings are chosen as the centroid of the embeddings of the contained elements. Our simulation easily realizes a logarithmic dependency between the MED and the number of elements to embed. These findings imply that embedding-based retrieval limitations stem primarily from learnability challenges, not geometric constraints, guiding future algorithm design.
Authors: Willams de Lima Costa, Thifany Ketuli Silva de Souza, Jonas Ferreira Silva, Carlos Gabriel Bezerra Pereira, Bruno Reis Vila Nova, Leonardo Silvino Brito, Rafael Raider Leoni, Juliano Silva, Valter Ferreira, Sibele Miguel Soares Neto, Samantha Uehara, Daniel Giacomo, Jo\~ao Marcelo Teixeira, Veronica Teichrieb, Cristiano Coelho de Ara\'ujo
Abstract: Road surface classification (RSC) is a key enabler for environment-aware predictive maintenance systems. However, existing RSC techniques often fail to generalize beyond narrow operational conditions due to limited sensing modalities and datasets that lack environmental diversity. This work addresses these limitations by introducing a multimodal framework that fuses images and inertial measurements using a lightweight bidirectional cross-attention module followed by an adaptive gating layer that adjusts modality contributions under domain shifts. Given the limitations of current benchmarks, especially regarding lack of variability, we introduce ROAD, a new dataset composed of three complementary subsets: (i) real-world multimodal recordings with RGB-IMU streams synchronized using a gold-standard industry datalogger, captured across diverse lighting, weather, and surface conditions; (ii) a large vision-only subset designed to assess robustness under adverse illumination and heterogeneous capture setups; and (iii) a synthetic subset generated to study out-of-distribution generalization in scenarios difficult to obtain in practice. Experiments show that our method achieves a +1.4 pp improvement over the previous state-of-the-art on the PVS benchmark and an +11.6 pp improvement on our multimodal ROAD subset, with consistently higher F1-scores on minority classes. The framework also demonstrates stable performance across challenging visual conditions, including nighttime, heavy rain, and mixed-surface transitions. These findings indicate that combining affordable camera and IMU sensors with multimodal attention mechanisms provides a scalable, robust foundation for road surface understanding, particularly relevant for regions where environmental variability and cost constraints limit the adoption of high-end sensing suites.
Authors: Weixin Chen, Li Chen, Yuhan Zhao
Abstract: Despite growing efforts to mitigate unfairness in recommender systems, existing fairness-aware methods typically fix the fairness requirement at training time and provide limited post-training flexibility. However, in real-world scenarios, diverse stakeholders may demand differing fairness requirements over time, so retraining for different fairness requirements becomes prohibitive. To address this limitation, we propose Cofair, a single-train framework that enables post-training fairness control in recommendation. Specifically, Cofair introduces a shared representation layer with fairness-conditioned adapter modules to produce user embeddings specialized for varied fairness levels, along with a user-level regularization term that guarantees user-wise monotonic fairness improvements across these levels. We theoretically establish that the adversarial objective of Cofair upper bounds demographic parity and the regularization term enforces progressive fairness at user level. Comprehensive experiments on multiple datasets and backbone models demonstrate that our framework provides dynamic fairness at different levels, delivering comparable or better fairness-accuracy curves than state-of-the-art baselines, without the need to retrain for each new fairness requirement. Our code is publicly available at https://github.com/weixinchen98/Cofair.
Authors: An\'ibal Silva, Mois\'es Santos, Andr\'e Restivo, Carlos Soares
Abstract: Tabular data remains a challenging domain for generative models. In particular, the standard Variational Autoencoder (VAE) architecture, typically composed of multilayer perceptrons, struggles to model relationships between features, especially when handling mixed data types. In contrast, Transformers, through their attention mechanism, are better suited for capturing complex feature interactions. In this paper, we empirically investigate the impact of integrating Transformers into different components of a VAE. We conduct experiments on 57 datasets from the OpenML CC18 suite and draw two main conclusions. First, results indicate that positioning Transformers to leverage latent and decoder representations leads to a trade-off between fidelity and diversity. Second, we observe a high similarity between consecutive blocks of a Transformer in all components. In particular, in the decoder, the relationship between the input and output of a Transformer is approximately linear.
Authors: Immanuel Abdi, Akshat Gupta, Micah Mok, Alexander Lu, Nicholas Lee, Gopala Anumanchipalli
Abstract: One of the biggest missing capabilities in current AI systems is the ability to learn continuously after deployment. Implementing such continually learning systems have several challenges, one of which is the large memory requirement of gradient-based algorithms that are used to train state-of-the-art LLMs. Evolutionary Strategies (ES) have recently re-emerged as a gradient-free alternative to traditional learning algorithms and have shown encouraging performance on specific tasks in LLMs. In this paper, we perform a comprehensive analysis of ES and specifically evaluate its forgetting curves when training for an increasing number of update steps. We first find that ES is able to reach performance numbers close to GRPO for math and reasoning tasks with a comparable compute budget. However, and most importantly for continual learning, the performance gains in ES is accompanied by significant forgetting of prior abilities, limiting its applicability for training models online. We also explore the reason behind this behavior and show that the updates made using ES are much less sparse and have orders of magnitude larger $\ell_2$ norm compared to corresponding GRPO updates, explaining the contrasting forgetting curves between the two algorithms. With this study, we aim to highlight the issue of forgetting in gradient-free algorithms like ES and hope to inspire future work to mitigate these issues.
Authors: Jingquan Wang, Andrew Negrut, Hongyu Wang, Harry Zhang, Dan Negrut
Abstract: We introduce SimBench, a benchmark designed to evaluate the proficiency of simulator-oriented LLMs (S-LLMs) in generating digital twins (DTs) that can be used in simulators for virtual testing. Given a collection of S-LLMs, this benchmark ranks them according to their ability to produce high-quality DTs. We demonstrate this by comparing over 33 open- and closed-source S-LLMs. Using multi-turn interactions, SimBench employs an LLM-as-a-judge (J-LLM) that leverages both predefined rules and human-in-the-loop guidance to assign scores for the DTs generated by the S-LLM, thus providing a consistent and expert-inspired evaluation protocol. The J-LLM is specific to a simulator, and herein the proposed benchmarking approach is demonstrated in conjunction with the open-sourceChrono multi-physics simulator. Chrono provided the backdrop used to assess an S-LLM in relation to the latter's ability to create digital twins for multibody dynamics, finite element analysis, vehicle dynamics, robotic dynamics, and sensor simulations. The proposed benchmarking principle is broadly applicable and enables the assessment of an S-LLM's ability to generate digital twins for other simulation packages, e.g., ANSYS, ABAQUS, OpenFOAM, StarCCM+, IsaacSim, and pyBullet.
Authors: Wenqing Zhou, Yuxuan Yan, Qianqian Yang
Abstract: Retrieval-Augmented Generation (RAG) improves factuality by grounding LLMs in external knowledge, yet conventional centralized RAG requires aggregating distributed data, raising privacy risks and incurring high retrieval latency and cost. We present DGRAG, a distributed graph-driven RAG framework for edge-cloud collaborative systems. Each edge device organizes local documents into a knowledge graph and periodically uploads subgraph-level summaries to the cloud for lightweight global indexing without exposing raw data. At inference time, queries are first answered on the edge; a gate mechanism assesses the confidence and consistency of multiple local generations to decide whether to return a local answer or escalate the query. For escalated queries, the cloud performs summary-based matching to identify relevant edges, retrieves supporting evidence from them, and generates the final response with a cloud LLM. Experiments on distributed question answering show that DGRAG consistently outperforms decentralized baselines while substantially reducing cloud overhead.
Authors: Florian Andreas Marwitz, Tanya Braun, Ralf M\"oller, Marcel Gehrke
Abstract: Decision making is a central problem in AI that can be formalized using a Markov Decision Process. A problem is that, with increasing numbers of (indistinguishable) objects, the state space grows exponentially. To compute policies, the state space has to be enumerated. Even more possibilities have to be enumerated if the size of the action space depends on the size of the state space, especially if we allow concurrent actions. To tackle the exponential blow-up in the action and state space, we present a first-order representation to store the spaces in polynomial instead of exponential size in the number of objects and introduce Foreplan, a relational forward planner, which uses this representation to efficiently compute policies for numerous indistinguishable objects and actions. Additionally, we introduce an even faster approximate version of Foreplan. Moreover, Foreplan identifies how many objects an agent should act on to achieve a certain task given restrictions. Further, we provide a theoretical analysis and an empirical evaluation of Foreplan, demonstrating a speedup of at least four orders of magnitude.
Authors: Kostis Michailidis, Dimos Tsouros, Tias Guns
Abstract: Discrete Combinatorial Problems (DCPs) are prevalent in industrial decision-making and optimisation. However, while constraint solving technologies for DCPs have advanced significantly, the core process of formalising them, namely constraint modelling, requires significant expertise and remains a bottleneck for wider adoption. Aiming to alleviate this bottleneck, recent studies have explored using Large Language Models (LLMs) to transform combinatorial problem descriptions into executable constraint models. However, the existing evaluation datasets for discrete constraint modelling are often limited to small, homogeneous, or domain-specific problems, which do not capture the diversity of real-world scenarios. This work addresses this gap by introducing DCP-Bench-Open, a novel benchmark that includes a diverse set of well-known discrete combinatorial problems sourced from the Constraint Programming (CP) and Operations Research (OR) communities, structured explicitly for evaluating LLM-driven constraint modelling. With this dataset, and given the variety of modelling frameworks, we compare and evaluate the modelling capabilities of LLMs for three distinct constraint modelling systems, which vary in abstraction level and underlying syntax. Notably, the results show higher performance when modelling with a high-level Python-based framework. Additionally, we systematically evaluate the use of prompt-based and inference-time compute methods across different LLMs, which further increase accuracy, reaching up to 91% on this highly challenging benchmark. DCP-Bench-Open is publicly available.
Authors: Fei Lin, Ziyang Gong, Cong Wang, Tengchao Zhang, Yonglin Tian, Yining Jiang, Ji Dai, Chao Guo, Xiaotong Yu, Xue Yang, Gen Luo, Fei-Yue Wang
Abstract: Toxicity remains a leading cause of early-stage drug development failure. Despite advances in molecular design and property prediction, the task of molecular toxicity repair, generating structurally valid molecular alternatives with reduced toxicity, has not yet been systematically defined or benchmarked. To fill this gap, we introduce ToxiMol, the first benchmark task for general-purpose Multimodal Large Language Models (MLLMs) focused on molecular toxicity repair. We construct a standardized dataset covering 11 primary tasks and 660 representative toxic molecules spanning diverse mechanisms and granularities. We design a prompt annotation pipeline with mechanism-aware and task-adaptive capabilities, informed by expert toxicological knowledge. In parallel, we propose an automated evaluation framework, ToxiEval, which integrates toxicity endpoint prediction, synthetic accessibility, drug-likeness, and structural similarity into a high-throughput evaluation chain for repair success. We systematically assess 43 mainstream general-purpose MLLMs and conduct multiple ablation studies to analyze key issues, including evaluation metrics, candidate diversity, and failure attribution. Experimental results show that although current MLLMs still face significant challenges on this task, they begin to demonstrate promising capabilities in toxicity understanding, semantic constraint adherence, and structure-aware editing.
Authors: Bohan Tang, Dezhao Luo, Jianheng Liu, Jingxuan Chen, Shaogang Gong, Jianye Hao, Jun Wang, Kun Shao
Abstract: The recent development of Large Language Models (LLMs) enables the rise of App agents that interpret user intent and operate smartphone Apps through actions such as clicking and scrolling. While prompt-based solutions with proprietary LLM APIs show promising ability, they incur heavy compute costs and external API dependency. Fine-tuning smaller open-source LLMs solves these limitations. However, current supervised fine-tuning methods use a syntax learning paradigm that forces agents to reproduce exactly the ground truth action strings, leading to out-of-distribution (OOD) vulnerability. To fill this gap, we propose Action Semantics Learning (ASL), a novel learning framework, where the learning objective is capturing the semantics of the ground truth actions. Specifically, inspired by the programming language theory, we define the action semantics for App agents as the state transition induced by the action in the user interface. Building on this insight, ASL employs a novel SEmantic Estimator~(SEE) to compute a semantic similarity to train the App agents in generating actions aligned with the semantics of ground truth actions, even when their syntactic forms differ. SEE is a flexible module that can be applied in both supervised and reinforcement fine-tuning paradigms. To support the effectiveness of ASL, we theoretically demonstrate the superior robustness of ASL for the OOD problem compared with the existing syntax learning paradigm. Extensive experiments across multiple offline and online benchmarks demonstrate that ASL significantly improves the accuracy and generalisation of App agents compared to existing methods.
Authors: Yi-Long Lu, Jiajun Song, Chunhui Zhang, Wei Wang
Abstract: Humans constantly generate a diverse range of tasks guided by internal motivations. While generative agents powered by large language models (LLMs) aim to simulate this complex behavior, it remains uncertain whether they operate on similar cognitive principles. To address this, we conducted a task-generation experiment comparing human responses with those of an LLM agent (GPT-4o). We find that human task generation is consistently influenced by psychological drivers, including personal values (e.g., Openness to Change) and cognitive style. Even when these psychological drivers are explicitly provided to the LLM, it fails to reflect the corresponding behavioral patterns. They produce tasks that are markedly less social, less physical, and thematically biased toward abstraction. Interestingly, while the LLM's tasks were perceived as more fun and novel, this highlights a disconnect between its linguistic proficiency and its capacity to generate human-like, embodied goals. We conclude that there is a core gap between the value-driven, embodied nature of human cognition and the statistical patterns of LLMs, highlighting the necessity of incorporating intrinsic motivation and physical grounding into the design of more human-aligned agents.
Authors: Wouter W. L. Nuijten, Mykola Lukashchuk, Thijs van de Laar, Bert de Vries
Abstract: We present a message passing approach to Expected Free Energy (EFE) minimization on factor graphs, based on the theory introduced in arXiv:2504.14898. By reformulating EFE minimization as Variational Free Energy minimization with epistemic priors, we transform a combinatorial search problem into a tractable inference problem solvable through standard variational techniques. Applying our message passing method to factorized state-space models enables efficient policy inference. We evaluate our method on environments with epistemic uncertainty: a stochastic gridworld and a partially observable Minigrid task. Agents using our approach consistently outperform conventional KL-control agents on these tasks, showing more robust planning and efficient exploration under uncertainty. In the stochastic gridworld environment, EFE-minimizing agents avoid risky paths, while in the partially observable minigrid setting, they conduct more systematic information-seeking. This approach bridges active inference theory with practical implementations, providing empirical evidence for the efficiency of epistemic priors in artificial agents.
Authors: Zakaria El Jaafari
Abstract: Monte Carlo Counterfactual Regret Minimization (MCCFR) has emerged as a cornerstone algorithm for solving extensive-form games, but its integration with deep neural networks introduces scale-dependent challenges that manifest differently across game complexities. This paper presents a comprehensive analysis of how neural MCCFR component effectiveness varies with game scale and proposes an adaptive framework for selective component deployment. We identify that theoretical risks such as nonstationary target distribution shifts, action support collapse, variance explosion, and warm-starting bias have scale-dependent manifestation patterns, requiring different mitigation strategies for small versus large games. Our proposed Robust Deep MCCFR framework incorporates target networks with delayed updates, uniform exploration mixing, variance-aware training objectives, and comprehensive diagnostic monitoring. Through systematic ablation studies on Kuhn and Leduc Poker, we demonstrate scale-dependent component effectiveness and identify critical component interactions. The best configuration achieves final exploitability of 0.0628 on Kuhn Poker, representing a 60% improvement over the classical framework (0.156). On the more complex Leduc Poker domain, selective component usage achieves exploitability of 0.2386, a 23.5% improvement over the classical framework (0.3703) and highlighting the importance of careful component selection over comprehensive mitigation. Our contributions include: (1) a formal theoretical analysis of risks in neural MCCFR, (2) a principled mitigation framework with convergence guarantees, (3) comprehensive multi-scale experimental validation revealing scale-dependent component interactions, and (4) practical guidelines for deployment in larger games.
Authors: Donghwan Lee, Hyukjun Yang, Bum Geun Park
Abstract: There are two primary approaches to solving Markov decision problems (MDPs): dynamic programming based on the Bellman equation and linear programming (LP). Dynamic programming methods are the most widely used and form the foundation of both classical and modern reinforcement learning (RL). By contrast, LP-based methods have been less commonly employed, although they have recently gained attention in contexts such as offline RL. The relative underuse of the LP-based methods stems from the fact that it leads to an inequality-constrained optimization problem, which is generally more challenging to solve effectively compared with Bellman-equation-based methods. The purpose of this paper is to establish a theoretical foundation for solving LP-based MDPs in a more effective and practical manner. Our key idea is to leverage the log-barrier function, widely used in inequality-constrained optimization, to transform the LP formulation of the MDP into an unconstrained optimization problem. This reformulation enables approximate solutions to be obtained easily via gradient descent. While the method may appear simple, to the best of our knowledge, a thorough theoretical interpretation of this approach has not yet been developed. This paper aims to bridge this gap.
Authors: Qian Cheng, Ruize Tang, Emilie Ma, Finn Hackett, Peiyang He, Yiming Su, Ivan Beschastnikh, Yu Huang, Xiaoxing Ma, Tianyin Xu
Abstract: Formal models are essential to specifying large, complex computer systems and verifying their correctness, but are notoriously expensive to write and maintain. Recent advances in generative AI show promise in generating certain forms of specifications. However, existing work mostly targets small code, not complete systems. It is unclear whether AI can deal with realistic system artifacts, as this requires abstracting their complex behavioral properties into formal models. We present SysMoBench, a benchmark that evaluates AI's ability to formally model large, complex systems. We focus on concurrent and distributed systems, which are keystones of today's critical computing infrastructures, encompassing operating systems and cloud infrastructure. We use TLA+, the de facto specification language for concurrent and distributed systems, though the benchmark can be extended to other specification languages. We address the primary challenge of evaluating AI-generated models by automating metrics like syntactic and runtime correctness, conformance to system code, and invariant correctness. SysMoBench currently includes eleven diverse system artifacts: the Raft implementation of Etcd and Redis, the leader election of ZooKeeper, the Spinlock, Mutex, and Ringbuffer in Asterinas OS, etc., with more being added. SysMoBench enables us to understand the capabilities and limitations of today's LLMs and agents, putting tools in this area on a firm footing and opening up promising new research directions.
Authors: Yunuo Cen, Zixuan Wang, Jintao Zhang, Zhiwei Zhang, Xuanyao Fong
Abstract: The Constraint-satisfaction problem (CSP) is fundamental in mathematics, physics, and theoretical computer science. Continuous local search (CLS) solvers, as recent advancements, can achieve highly competitive results on certain classes of Boolean satisfiability (SAT) problems. Motivated by these advances, we extend the CLS framework from Boolean SAT to general CSP with finite-domain variables and expressive constraint formulations. We present FourierCSP, a continuous optimization framework that generalizes the Walsh-Fourier transform to CSP, allowing for transforming versatile constraints to compact multilinear polynomials, thereby avoiding the need for auxiliary variables and memory-intensive encodings. We employ projected subgradient and mirror descent algorithms with provable convergence guarantees, and further combine them to accelerate gradient-based optimization. Empirical results on benchmark suites demonstrate that FourierCSP is scalable and competitive, significantly broadening the class of problems that can be efficiently solved by differentiable CLS techniques and paving the way toward end-to-end neurosymbolic integration.
Authors: Chen Li, Zhantao Yang, Han Zhang, Fangyi Chen, Chenchen Zhu, Anudeepsekhar Bolimera, Marios Savvides
Abstract: Vision-Language-Action (VLA) models show promise in embodied reasoning, yet remain far from true generalists-they often require task-specific fine-tuning, incur high compute costs, and generalize poorly to unseen tasks. We propose MetaVLA, a unified, backbone-agnostic post-training framework for efficient and scalable alignment. MetaVLA introduces Context-Aware Meta Co-Training, which consolidates diverse target tasks into a single fine-tuning stage while leveraging structurally diverse auxiliary tasks to improve in-domain generalization. Unlike naive multi-task SFT, MetaVLA integrates a lightweight meta-learning mechanism-derived from Attentive Neural Processes-to enable rapid adaptation from diverse contexts with minimal architectural change or inference overhead. On the LIBERO benchmark, MetaVLA with six auxiliary tasks outperforms OpenVLA by up to 8.0% on long-horizon tasks, reduces training steps from 240K to 75K, and cuts GPU time by ~76%. These results show that scalable, low-resource post-training is achievable-paving the way toward general-purpose embodied agents. Code will be available.
Authors: Pedro Antonio Alarcon Granadeno, Arturo Miguel Bernal Russell, Sofia Nelson, Demetrius Hernandez, Maureen Petterson, Michael Murphy, Walter J. Scheirer, Jane Cleland-Huang
Abstract: Cyber-physical systems increasingly rely on Foundational Models such as Large Language Models (LLMs) and Vision-Language Models (VLMs) to increase autonomy through enhanced perception, inference, and planning. However, these models also introduce new types of errors, such as hallucinations, overgeneralizations, and context misalignments, resulting in incorrect and flawed decisions. To address this, we introduce the concept of Cognition Envelopes, designed to establish reasoning boundaries that constrain AI-generated decisions while complementing the use of meta-cognition and traditional safety envelopes. As with safety envelopes, Cognition Envelopes require practical guidelines and systematic processes for their definition, validation, and assurance.
Authors: Yang You, Ufuk \c{C}ak{\i}r, Alex Schutz, Nick Hawes
Abstract: The value function of a POMDP exhibits the piecewise-linear-convex (PWLC) property and can be represented as a finite set of hyperplanes, known as $\alpha$-vectors. Most state-of-the-art POMDP solvers (offline planners) follow the point-based value iteration scheme, which performs Bellman backups on $\alpha$-vectors at reachable belief points until convergence. However, since each $\alpha$-vector is $|S|$-dimensional, these methods quickly become intractable for large-scale problems due to the prohibitive computational cost of Bellman backups. In this work, we demonstrate that the PWLC property allows a POMDP's value function to be alternatively represented as a finite set of neural networks. This insight enables a novel POMDP planning algorithm called \emph{Neural Value Iteration}, which combines the generalization capability of neural networks with the classical value iteration framework. Our approach achieves near-optimal solutions even in extremely large POMDPs that are intractable for existing offline solvers.
Authors: Bakhtawar Ahtisham, Kirk Vanacore, Jinsook Lee, Zhuqian Zhou, Doug Pietrzak, Rene F. Kizilcec
Abstract: Large Language Models (LLMs) are increasingly used to annotate learning interactions, yet concerns about reliability limit their utility. We test whether verification-oriented orchestration-prompting models to check their own labels (self-verification) or audit one another (cross-verification)-improves qualitative coding of tutoring discourse. Using transcripts from 30 one-to-one math sessions, we compare three production LLMs (GPT, Claude, Gemini) under three conditions: unverified annotation, self-verification, and cross-verification across all orchestration configurations. Outputs are benchmarked against a blinded, disagreement-focused human adjudication using Cohen's kappa. Overall, orchestration yields a 58 percent improvement in kappa. Self-verification nearly doubles agreement relative to unverified baselines, with the largest gains for challenging tutor moves. Cross-verification achieves a 37 percent improvement on average, with pair- and construct-dependent effects: some verifier-annotator pairs exceed self-verification, while others reduce alignment, reflecting differences in verifier strictness. We contribute: (1) a flexible orchestration framework instantiating control, self-, and cross-verification; (2) an empirical comparison across frontier LLMs on authentic tutoring data with blinded human "gold" labels; and (3) a concise notation, verifier(annotator) (e.g., Gemini(GPT) or Claude(Claude)), to standardize reporting and make directional effects explicit for replication. Results position verification as a principled design lever for reliable, scalable LLM-assisted annotation in Learning Analytics.
Authors: Danial Safaei, Siddartha Khastgir, Mohsen Alirezaei, Jeroen Ploeg, Son Tong, Chih-Hong Cheng, Xingyu Zhao
Abstract: Virtual testing using synthetic data has become a cornerstone of autonomous vehicle (AV) safety assurance. Despite progress in improving visual realism through advanced simulators and generative AI, recent studies reveal that pixel-level fidelity alone does not ensure reliable transfer from simulation to the real world. What truly matters is whether the system-under-test (SUT) bases its decisions on consistent decision evidence in both real and simulated environments, not just whether images "look real" to humans. To this end this paper proposes a behavior-grounded fidelity measure by introducing Decisive Feature Fidelity (DFF), a new SUT-specific metric that extends the existing fidelity spectrum to capture mechanism parity, that is, agreement in the model-specific decisive evidence that drives the SUT's decisions across domains. DFF leverages explainable-AI methods to identify and compare the decisive features driving the SUT's outputs for matched real-synthetic pairs. We further propose estimators based on counterfactual explanations, along with a DFF-guided calibration scheme to enhance simulator fidelity. Experiments on 2126 matched KITTI-VirtualKITTI2 pairs demonstrate that DFF reveals discrepancies overlooked by conventional output-value fidelity. Furthermore, results show that DFF-guided calibration improves decisive-feature and input-level fidelity without sacrificing output value fidelity across diverse SUTs.
Authors: G\"okdeniz G\"ulmez
Abstract: We present Gabliteration, a novel neural weight modification technique that advances beyond traditional abliteration methods by implementing adaptive multi-directional projections with regularized layer selection. Our approach addresses the fundamental limitation of existing methods that compromise model quality while attempting to modify specific behavioral patterns. Through dynamic layer optimization, regularized projection matrices, and adaptive scaling mechanisms, we achieve theoretically superior weight modification while minimizing quality degradation in unrelated domains. We validate our method through the gabliterated-v1 model series (0.6B to 4B parameters) available on Hugging Face, demonstrating practical applicability across multiple model scales.
Authors: Xu Huang, Junwu Chen, Yuxing Fei, Zhuohan Li, Philippe Schwaller, Gerbrand Ceder
Abstract: Large language model (LLM) agents currently depend on predefined tools or early-stage tool generation, limiting their adaptability and scalability to complex scientific tasks. We introduce CASCADE, a self-evolving agentic framework representing an early instantiation of the transition from "LLM + tool use" to "LLM + skill acquisition". CASCADE enables agents to master complex external tools and codify knowledge through two meta-skills: continuous learning via web search, code extraction, and memory utilization; self-reflection via introspection, knowledge graph exploration, and others. We evaluate CASCADE on SciSkillBench, a benchmark of 116 materials science and chemistry research tasks. CASCADE achieves a 93.3% success rate using GPT-5, compared to 35.4% without evolution mechanisms. We further demonstrate real-world applications in computational analysis, autonomous laboratory experiments, and selective reproduction of published papers. Along with human-agent collaboration and memory consolidation, CASCADE accumulates executable skills that can be shared across agents and scientists, moving toward scalable AI-assisted scientific research.
Authors: Alex L. Zhang, Tim Kraska, Omar Khattab
Abstract: We study allowing large language models (LLMs) to process arbitrarily long prompts through the lens of inference-time scaling. We propose Recursive Language Models (RLMs), a general inference paradigm that treats long prompts as part of an external environment and allows the LLM to programmatically examine, decompose, and recursively call itself over snippets of the prompt. We find that RLMs can successfully process inputs up to two orders of magnitude beyond model context windows and, even for shorter prompts, dramatically outperform the quality of vanilla frontier LLMs and common long-context scaffolds across four diverse long-context tasks while having comparable cost. At a small scale, we post-train the first natively recursive language model. Our model, RLM-Qwen3-8B, outperforms the underlying Qwen3-8B model by $28.3\%$ on average and even approaches the quality of vanilla GPT-5 on three long-context tasks. Code is available at https://github.com/alexzhang13/rlm.
Authors: Jiaqi Liu, Yaofeng Su, Peng Xia, Siwei Han, Zeyu Zheng, Cihang Xie, Mingyu Ding, Huaxiu Yao
Abstract: To support long-term interaction in complex environments, LLM agents require memory systems that manage historical experiences. Existing approaches either retain full interaction histories via passive context extension, leading to substantial redundancy, or rely on iterative reasoning to filter noise, incurring high token costs. To address this challenge, we introduce SimpleMem, an efficient memory framework based on semantic lossless compression. We propose a three-stage pipeline designed to maximize information density and token utilization: (1) Semantic Structured Compression, which distills unstructured interactions into compact, multi-view indexed memory units; (2) Online Semantic Synthesis, an intra-session process that instantly integrates related context into unified abstract representations to eliminate redundancy; and (3) Intent-Aware Retrieval Planning, which infers search intent to dynamically determine retrieval scope and construct precise context efficiently. Experiments on benchmark datasets show that our method consistently outperforms baseline approaches in accuracy, retrieval efficiency, and inference cost, achieving an average F1 improvement of 26.4% while reducing inference-time token consumption by up to 30-fold, demonstrating a superior balance between performance and efficiency. Code is available at https://github.com/aiming-lab/SimpleMem.
Authors: Sunzhu Li, Jiale Zhao, Miteto Wei, Huimin Ren, Yang Zhou, Jingwen Yang, Shunyu Liu, Kaike Zhang, Wei Chen
Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has driven substantial progress in reasoning-intensive domains like mathematics. However, optimizing open-ended generation remains challenging due to the lack of ground truth. While rubric-based evaluation offers a structured proxy for verification, existing methods suffer from scalability bottlenecks and coarse criteria, resulting in a supervision ceiling effect. To address this, we propose an automated Coarse-to-Fine Rubric Generation framework. By synergizing principle-guided synthesis, multi-model aggregation, and difficulty evolution, our approach produces comprehensive and highly discriminative criteria capable of capturing the subtle nuances. Based on this framework, we introduce RubricHub, a large-scale ($\sim$110k) and multi-domain dataset. We validate its utility through a two-stage post-training pipeline comprising Rubric-based Rejection Sampling Fine-Tuning (RuFT) and Reinforcement Learning (RuRL). Experimental results demonstrate that RubricHub unlocks significant performance gains: our post-trained Qwen3-14B achieves state-of-the-art (SOTA) results on HealthBench (69.3), surpassing proprietary frontier models such as GPT-5. Our code is available at \href{https://github.com/teqkilla/RubricHub}{ this URL}.
Authors: Xinyu Zhu, Yuzhu Cai, Zexi Liu, Bingyang Zheng, Cheng Wang, Rui Ye, Jiaao Chen, Hanrui Wang, Wei-Chen Wang, Yuzhi Zhang, Linfeng Zhang, Weinan E, Di Jin, Siheng Chen, Yanfeng Wang
Abstract: The advancement of artificial intelligence toward agentic science is currently bottlenecked by the challenge of ultra-long-horizon autonomy, the ability to sustain strategic coherence and iterative correction over experimental cycles spanning days or weeks. While Large Language Models (LLMs) have demonstrated prowess in short-horizon reasoning, they are easily overwhelmed by execution details in the high-dimensional, delayed-feedback environments of real-world research, failing to consolidate sparse feedback into coherent long-term guidance. Here, we present ML-Master 2.0, an autonomous agent that masters ultra-long-horizon machine learning engineering (MLE) which is a representative microcosm of scientific discovery. By reframing context management as a process of cognitive accumulation, our approach introduces Hierarchical Cognitive Caching (HCC), a multi-tiered architecture inspired by computer systems that enables the structural differentiation of experience over time. By dynamically distilling transient execution traces into stable knowledge and cross-task wisdom, HCC allows agents to decouple immediate execution from long-term experimental strategy, effectively overcoming the scaling limits of static context windows. In evaluations on OpenAI's MLE-Bench under 24-hour budgets, ML-Master 2.0 achieves a state-of-the-art medal rate of 56.44%. Our findings demonstrate that ultra-long-horizon autonomy provides a scalable blueprint for AI capable of autonomous exploration beyond human-precedent complexities.
Authors: Pietro Barbiero, Mateo Espinosa Zarlenga, Francesco Giannini, Alberto Termine, Filippo Bonchi, Mateja Jamnik, Giuseppe Marra
Abstract: This paper argues that interpretability research in Artificial Intelligence (AI) is fundamentally ill-posed as existing definitions of interpretability fail to describe how interpretability can be formally tested or designed for. We posit that actionable definitions of interpretability must be formulated in terms of *symmetries* that inform model design and lead to testable conditions. Under a probabilistic view, we hypothesise that four symmetries (inference equivariance, information invariance, concept-closure invariance, and structural invariance) suffice to (i) formalise interpretable models as a subclass of probabilistic models, (ii) yield a unified formulation of interpretable inference (e.g., alignment, interventions, and counterfactuals) as a form of Bayesian inversion, and (iii) provide a formal framework to verify compliance with safety standards and regulations.
Authors: Michele Loi
Abstract: Large language models increasingly function as artificial reasoners: they evaluate arguments, assign credibility, and express confidence. Yet their belief-forming behavior is governed by implicit, uninspected epistemic policies. This paper argues for an epistemic constitution for AI: explicit, contestable meta-norms that regulate how systems form and express beliefs. Source attribution bias provides the motivating case: I show that frontier models enforce identity-stance coherence, penalizing arguments attributed to sources whose expected ideological position conflicts with the argument's content. When models detect systematic testing, these effects collapse, revealing that systems treat source-sensitivity as bias to suppress rather than as a capacity to execute well. I distinguish two constitutional approaches: the Platonic, which mandates formal correctness and default source-independence from a privileged standpoint, and the Liberal, which refuses such privilege, specifying procedural norms that protect conditions for collective inquiry while allowing principled source-attending grounded in epistemic vigilance. I argue for the Liberal approach, sketch a constitutional core of eight principles and four orientations, and propose that AI epistemic governance requires the same explicit, contestable structure we now expect for AI ethics.
Authors: Mingyang Song, Haoyu Sun, Jiawei Gu, Linjie Li, Luxin Xu, Ranjay Krishna, Yu Cheng
Abstract: When humans face problems beyond their immediate capabilities, they rely on tools, providing a promising paradigm for improving visual reasoning in multimodal large language models (MLLMs). Effective reasoning, therefore, hinges on knowing which tools to use, when to invoke them, and how to compose them over multiple steps, even when faced with new tools or new tasks. We introduce \textbf{AdaReasoner}, a family of multimodal models that learn tool use as a general reasoning skill rather than as tool-specific or explicitly supervised behavior. AdaReasoner is enabled by (i) a scalable data curation pipeline exposing models to long-horizon, multi-step tool interactions; (ii) Tool-GRPO, a reinforcement learning algorithm that optimizes tool selection and sequencing based on end-task success; and (iii) an adaptive learning mechanism that dynamically regulates tool usage. Together, these components allow models to infer tool utility from task context and intermediate outcomes, enabling coordination of multiple tools and generalization to unseen tools. Empirically, AdaReasoner exhibits strong tool-adaptive and generalization behaviors: it autonomously adopts beneficial tools, suppresses irrelevant ones, and adjusts tool usage frequency based on task demands, despite never being explicitly trained to do so. These capabilities translate into state-of-the-art performance across challenging benchmarks, improving the 7B base model by +24.9\% on average and surpassing strong proprietary systems such as GPT-5 on multiple tasks, including VSP and Jigsaw.
Authors: Qiyuan Xu, Xiaokun Luan, Renxi Wang, Joshua Ong Jun Leang, Peixin Wang, Haonan Li, Wenda Li, Conrad Watt
Abstract: Theorem proving is fundamental to program verification, where the automated proof of Verification Conditions (VCs) remains a primary bottleneck. Real-world program verification frequently encounters hard VCs that existing Automated Theorem Provers (ATPs) cannot prove, leading to a critical need for extensive manual proofs that burden practical application. While Neural Theorem Proving (NTP) has achieved significant success in mathematical competitions, demonstrating the potential of machine learning approaches to formal reasoning, its application to program verification--particularly VC proving--remains largely unexplored. Despite existing work on annotation synthesis and verification-related theorem proving, no benchmark has specifically targeted this fundamental bottleneck: automated VC proving. This work introduces Neural Theorem Proving for Verification Conditions (NTP4VC), presenting the first real-world multi-language benchmark for this task. From real-world projects such as Linux and Contiki-OS kernel, our benchmark leverages industrial pipelines (Why3 and Frama-C) to generate semantically equivalent test cases across formal languages of Isabelle, Lean, and Rocq. We evaluate large language models (LLMs), both general-purpose and those fine-tuned for theorem proving, on NTP4VC. Results indicate that although LLMs show promise in VC proving, significant challenges remain for program verification, highlighting a large gap and opportunity for future research.
Authors: Young In Kim, Andrea Agiollo, Pratiksha Agrawal, Johannes O. Royset, Rajiv Khanna
Abstract: Optimization algorithms that seek flatter minima, such as Sharpness-Aware Minimization (SAM), are credited with improved generalization and robustness to noise. We ask whether such gains impact membership privacy. Surprisingly, we find that SAM is more prone to Membership Inference Attacks (MIA) than classical SGD across multiple datasets and attack methods, despite achieving lower test error. This suggests that the geometric mechanism of SAM that improves generalization simultaneously exacerbates membership leakage. We investigate this phenomenon through extensive analysis of memorization and influence scores. Our results reveal that SAM is more capable of capturing atypical subpatterns, leading to higher memorization scores of samples. Conversely, SGD depends more heavily on majority features, exhibiting worse generalization on atypical subgroups and lower memorization. Crucially, this characteristic of SAM can be linked to lower variance in the prediction confidence of unseen samples, thereby amplifying membership signals. Finally, we model SAM under a perfectly interpolating linear regime and theoretically show that sharpness regularization inherently reduces variance, guaranteeing a higher MIA advantage for confidence and likelihood ratio attacks.
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: Shanshan Han, Qifan Zhang, Weizhao Jin, Zhaozhuo Xu
Abstract: This paper explores multi-agent systems and identify challenges that remain inadequately addressed. By leveraging the diverse capabilities and roles of individual agents, multi-agent systems can tackle complex tasks through agent collaboration. We discuss optimizing task allocation, fostering robust reasoning through iterative debates, managing complex and layered context information, and enhancing memory management to support the intricate interactions within multi-agent systems. We also explore potential applications of multi-agent systems in blockchain systems to shed light on their future development and application in real-world distributed systems.
Authors: Zeming Dong, Qiang Hu, Xiaofei Xie, Maxime Cordy, Mike Papadakis, Yves Le Traon, Jianjun Zhao
Abstract: Pre-trained code models lead the era of code intelligence, with multiple models designed with impressive performance. However, one important problem, data augmentation for code data that automatically helps developers prepare training data lacks study in this field. In this paper, we introduce a generic data augmentation framework, GenCode, to enhance the training of code understanding models. Simply speaking, GenCode follows a generation-and-selection paradigm to prepare useful training code data. Specifically, it employs code augmentation techniques to generate new code candidates first and then identifies important ones as the training data by influence scores. To evaluate the effectiveness of GenCode, we conduct experiments on four code understanding tasks (e.g., code clone detection) and three pre-trained code models (e.g., CodeT5) and two recent released code-specific Large Language Models (LLMs) (e.g., Qwen2.5-Coder). Compared to the state-of-the-art (SOTA) code augmentation method MixCode, GenCode produces pre-trained code models with 2.92% higher accuracy and 4.90% adversarial robustness on average. For code-specific LLMs, GenCode achieves an average improvement of 0.93% in accuracy and 0.98% in natural robustness.
Authors: Dylan Bouchard
Abstract: Bias and fairness risks in Large Language Models (LLMs) vary substantially across deployment contexts, yet existing approaches lack systematic guidance for selecting appropriate evaluation metrics. We present a decision framework that maps LLM use cases, characterized by a model and population of prompts, to relevant bias and fairness metrics based on task type, whether prompts contain protected attribute mentions, and stakeholder priorities. Our framework addresses toxicity, stereotyping, counterfactual unfairness, and allocational harms, and introduces novel metrics based on stereotype classifiers and counterfactual adaptations of text similarity measures. All metrics require only LLM outputs for computation, simplifying implementation while avoiding embedding-based approaches that often correlate poorly with downstream harms. We provide an open-source Python library, LangFair, for practical adoption. Extensive experiments demonstrate that fairness risks cannot be reliably assessed from benchmark performance alone: results on one prompt dataset likely overstate or understate risks for another, underscoring that fairness evaluation must be grounded in the specific deployment context.
Authors: Danlu Chen, Freda Shi, Aditi Agarwal, Jacobo Myerston, Taylor Berg-Kirkpatrick
Abstract: Standard natural language processing (NLP) pipelines operate on symbolic representations of language, which typically consist of sequences of discrete tokens. However, creating an analogous representation for ancient logographic writing systems is an extremely labor intensive process that requires expert knowledge. At present, a large portion of logographic data persists in a purely visual form due to the absence of transcription -- this issue poses a bottleneck for researchers seeking to apply NLP toolkits to study ancient logographic languages: most of the relevant data are images of writing. This paper investigates whether direct processing of visual representations of language offers a potential solution. We introduce LogogramNLP, the first benchmark enabling NLP analysis of ancient logographic languages, featuring both transcribed and visual datasets for four writing systems along with annotations for tasks like classification, translation, and parsing. Our experiments compare systems that employ recent visual and text encoding strategies as backbones. The results demonstrate that visual representations outperform textual representations for some investigated tasks, suggesting that visual processing pipelines may unlock a large amount of cultural heritage data of logographic languages for NLP-based analyses.
Authors: Vincent Freiberger, Arthur Fleig, Erik Buchmann
Abstract: Understanding and engaging with privacy policies is crucial for online privacy, yet these documents remain notoriously complex and difficult to navigate. We present PRISMe, an interactive browser extension that combines LLM-based policy assessment with a dashboard and customizable chat interface, enabling users to skim quick overviews or explore policy details in depth while browsing. We conduct a user study (N=22) with participants of diverse privacy knowledge to investigate how users interpret the tool's explanations and how it shapes their engagement with privacy policies, identifying distinct interaction patterns. Participants valued the clear overviews and conversational depth, but flagged some issues, particularly adversarial robustness and hallucination risks. Thus, we investigate how a retrieval-augmented generation (RAG) approach can alleviate issues by re-running the chat queries from the study. Our findings surface design challenges as well as technical trade-offs, contributing actionable insights for developing future user-centered, trustworthy privacy policy analysis tools.
Authors: Xianglin Yang, Gelei Deng, Jieming Shi, Tianwei Zhang, Jin Song Dong
Abstract: Large language models (LLMs) are vital for a wide range of applications yet remain susceptible to jailbreak threats, which could lead to the generation of inappropriate responses. Conventional defenses, such as refusal and adversarial training, often fail to cover corner cases or rare domains, leaving LLMs still vulnerable to more sophisticated attacks. We propose a novel defense strategy, Safety Chain-of-Thought (SCoT), which harnesses the enhanced \textit{reasoning capabilities} of LLMs for proactive assessment of harmful inputs, rather than simply blocking them. SCoT augments any refusal training datasets to critically analyze the intent behind each request before generating answers. By employing proactive reasoning, SCoT enhances the generalization of LLMs across varied harmful queries and scenarios not covered in the safety alignment corpus. Additionally, it generates detailed refusals specifying the rules violated. Comparative evaluations show that SCoT significantly surpasses existing defenses, reducing vulnerability to out-of-distribution issues and adversarial manipulations while maintaining strong general capabilities.
Authors: Shanzhi Gu, Zhaoyang Qu, Ruotong Geng, Mingyang Geng, Shangwen Wang, Chuanfu Xu, Haotian Wang, Zhipeng Lin, Dezun Dong
Abstract: Large Language Models for Code (LLMs4Code) have achieved strong performance in code generation, but recent studies reveal that they may memorize and leak sensitive information contained in training data, posing serious privacy risks. To address this gap, this work presents the first comprehensive empirical study on applying machine unlearning to mitigate sensitive information leakage in LLMs4Code. We first construct a dedicated benchmark that includes: (i) a synthetic forget set containing diverse forms of personal information, and (ii) a retain set designed to evaluate whether code-generation capability is preserved after unlearning. Using this benchmark, we systematically assess three representative unlearning algorithms (GA, GA+GD, GA+KL) across three widely used open-source LLMs4Code models (AIXCoder-7B, CodeLlama-7B, CodeQwen-7B). Experimental results demonstrate that machine unlearning can substantially reduce direct memorization-based leakage: on average, the direct leak rate drops by more than 50% while retaining about over 91% of the original code-generation performance. Moreover, by analyzing post-unlearning outputs, we uncover a consistent shift from direct to indirect leakage, revealing an underexplored vulnerability that persists even when the target data has been successfully forgotten. Our findings show that machine unlearning is a feasible and effective solution for enhancing privacy protection in LLMs4Code, while also highlighting the need for future techniques capable of mitigating both direct and indirect leakage simultaneously.
Authors: Ilan Lobel, Humberto Moreira, Omar Mouchtaki
Abstract: We study auction design when a seller relies on machine-learning predictions of bidders' valuations that may be unreliable. Motivated by modern ML systems that are often accurate but occasionally fail in a way that is essentially uninformative, we model predictions as randomly wrong: with high probability the signal equals the bidder's true value, and otherwise it is a hallucination independent of the value. We analyze revenue-maximizing auctions when the seller publicly reveals these signals. A central difficulty is that the resulting posterior belief combines a continuous distribution with a point mass at the signal, so standard Myerson techniques do not directly apply. We provide a tractable characterization of the optimal signal-revealing auction by providing a closed-form characterization of the appropriate ironed virtual values. This characterization yields simple and intuitive implications. With a single bidder, the optimal mechanism reduces to a posted-price policy with a small number of regimes: the seller ignores low signals, follows intermediate signals, caps moderately high signals, and may again follow very high signals. With multiple bidders, we show that a simple eager second-price auction with signal-dependent reserve prices performs nearly optimally in numerical experiments and substantially outperforms natural benchmarks that either ignore the signal or treat it as fully reliable.
Authors: Haonan Yu, Junhao Liu, Xin Zhang
Abstract: Anchors is a popular local model-agnostic explanation technique whose applicability is limited by its computational inefficiency. To address this limitation, we propose a memorization-based framework that accelerates Anchors while preserving explanation fidelity and interpretability. Our approach leverages the iterative nature of Anchors' algorithm which gradually refines an explanation until it is precise enough for a given input by storing and reusing intermediate results obtained during prior explanations. Specifically, we maintain a memory of low-precision, high-coverage rules and introduce a rule transformation framework to adapt them to new inputs: the horizontal transformation adapts a pre-trained explanation to the current input by replacing features, and the vertical transformation refines the general explanation until it is precise enough for the input. We evaluate our method across tabular, text, and image datasets, demonstrating that it significantly reduces explanation generation time while maintaining fidelity and interpretability, thereby enabling the practical adoption of Anchors in time-sensitive applications.
Authors: Yiyang Lu, Mohammad Pedramfar, Vaneet Aggarwal
Abstract: Projection-based algorithms for Constrained Online Convex Optimization (COCO) achieve optimal $\mathcal{O}(T^{1/2})$ regret guarantees but face scalability challenges due to the computational complexity of projections. To circumvent this, projection-free methods utilizing Linear Optimization Oracles (LOO) have been proposed, albeit typically achieving slower $\mathcal{O}(T^{3/4})$ regret rates. In this work, we examine whether the $\mathcal{O}(T^{1/2})$ rate can be recovered in the projection-free setting by strengthening the oracle assumption. We introduce BAGEL, an algorithm utilizing a Separation Oracle (SO) that achieves $\mathcal{O}(T^{1/2})$ regret and $\tilde{\mathcal{O}}(T^{1/2})$ cumulative constraint violation (CCV) for convex cost functions. Our analysis shows that by leveraging an infeasible projection via SO, we can match the time-horizon dependence of projection-based methods with $\tilde{\mathcal{O}}(T)$ oracle calls, provided dependence on the geometry of the action set. This establishes a specific regime where projection-free methods can attain the same convergence rates as projection-based counterparts.
Authors: Gilad Yehudai, Noah Amsel, Joan Bruna
Abstract: It is well understood that different neural network architectures are suited to different tasks, but is there always a single best architecture for a given task? We compare the expressive power of transformers, RNNs, and transformers with chain of thought tokens on a simple and natural class of tasks we term Compositional Reasoning Questions (CRQ). This family captures multi-step problems with tree-like compositional structure, such as evaluating Boolean formulas. We prove that under standard hardness assumptions, \emph{none} of these three architectures is capable of solving CRQs unless some hyperparameter (depth, embedding dimension, and number of chain of thought tokens, respectively) grows with the size of the input. We then provide constructions for solving CRQs with each architecture. For transformers, our construction uses depth that is logarithmic in the problem size. For RNNs, logarithmic embedding dimension is necessary and sufficient, so long as the inputs are provided in a certain order. For transformers with chain of thought, our construction uses $n$ CoT tokens for input size $n$. These results show that, while CRQs are inherently hard, there are several different ways for language models to overcome this hardness. Even for a single class of problems, each architecture has strengths and weaknesses, and none is strictly better than the others.
Authors: Gilad Yehudai, Clayton Sanford, Maya Bechler-Speicher, Orr Fischer, Ran Gilad-Bachrach, Amir Globerson
Abstract: Transformers have revolutionized the field of machine learning. In particular, they can be used to solve complex algorithmic problems, including graph-based tasks. In such algorithmic tasks a key question is what is the minimal size of a transformer that can implement the task. Recent work has begun to explore this problem for graph-based tasks, showing that for sub-linear embedding dimension (i.e., model width) logarithmic depth suffices. However, an open question, which we address here, is what happens if width is allowed to grow linearly, while depth is kept fixed. Here we analyze this setting, and provide the surprising result that with linear width, constant depth suffices for solving a host of graph-based problems. This suggests that a moderate increase in width can allow much shallower models, which are advantageous in terms of inference and train time. For other problems, we show that quadratic width is required. Our results demonstrate the complex and intriguing landscape of transformer implementations of graph-based algorithms. We empirically investigate these trade-offs between the relative powers of depth and width and find tasks where wider models have the same accuracy as deep models, while having much faster train and inference time due to parallelizable hardware.
Authors: David Bani-Harouni, Chantal Pellegrini, Paul Stangel, Ege \"Ozsoy, Kamilia Zaripova, Matthias Keicher, Nassir Navab
Abstract: A safe and trustworthy use of Large Language Models (LLMs) requires an accurate expression of confidence in their answers. We propose a novel Reinforcement Learning approach that allows to directly fine-tune LLMs to express calibrated confidence estimates alongside their answers to factual questions. Our method optimizes a reward based on the logarithmic scoring rule, explicitly penalizing both over- and under-confidence. This encourages the model to align its confidence estimates with the actual predictive accuracy. The optimal policy under our reward design would result in perfectly calibrated confidence expressions. Unlike prior approaches that decouple confidence estimation from response generation, our method integrates confidence calibration seamlessly into the generative process of the LLM. Empirically, we demonstrate that models trained with our approach exhibit substantially improved calibration and generalize to unseen tasks without further fine-tuning, suggesting the emergence of general confidence awareness.
Authors: Haohao Qu, Shanru Lin, Yujuan Ding, Yiqi Wang, Wenqi Fan
Abstract: Recent advances in generative artificial intelligence, particularly large language models (LLMs), have opened new opportunities for enhancing recommender systems (RecSys). Most existing LLM-based RecSys approaches operate in a discrete space, using vector-quantized tokenizers to align with the inherent discrete nature of language models. However, these quantization methods often result in lossy tokenization and suboptimal learning, primarily due to inaccurate gradient propagation caused by the non-differentiable argmin operation in standard vector quantization. Inspired by the emerging trend of embracing continuous tokens in language models, we propose ContRec, a novel framework that seamlessly integrates continuous tokens into LLM-based RecSys. Specifically, ContRec consists of two key modules: a sigma-VAE Tokenizer, which encodes users/items with continuous tokens; and a Dispersive Diffusion module, which captures implicit user preference. The tokenizer is trained with a continuous Variational Auto-Encoder (VAE) objective, where three effective techniques are adopted to avoid representation collapse. By conditioning on the previously generated tokens of the LLM backbone during user modeling, the Dispersive Diffusion module performs a conditional diffusion process with a novel Dispersive Loss, enabling high-quality user preference generation through next-token diffusion. Finally, ContRec leverages both the textual reasoning output from the LLM and the latent representations produced by the diffusion model for Top-K item retrieval, thereby delivering comprehensive recommendation results. Extensive experiments on four datasets demonstrate that ContRec consistently outperforms both traditional and SOTA LLM-based recommender systems. Our results highlight the potential of continuous tokenization and generative modeling for advancing the next generation of recommender systems.
Authors: Jing Han, Hanting Chen, Kai Han, Xiaomeng Huang, Wenjun Xu, Dacheng Tao, Ping Zhang
Abstract: This position paper argues that the next generation of artificial intelligence in meteorological and climate sciences must transition from fragmented hybrid heuristics toward a unified paradigm of physics-guided multimodal transformers. While purely data-driven models have achieved significant gains in predictive accuracy, they often treat atmospheric processes as mere visual patterns, frequently producing results that lack scientific consistency or violate fundamental physical laws. We contend that current ``hybrid'' attempts to bridge this gap remain ad-hoc and struggle to scale across the heterogeneous nature of meteorological data ranging from satellite imagery to sparse sensor measurements. We argue that the transformer architecture, through its inherent capacity for cross-modal alignment, provides the only viable foundation for a systematic integration of domain knowledge via physical constraint embedding and physics-informed loss functions. By advocating for this unified architectural shift, we aim to steer the community away from ``black-box'' curve fitting and toward AI systems that are inherently falsifiable, scientifically grounded, and robust enough to address the existential challenges of extreme weather and climate change.
Authors: Lawrence Liu, Inesh Chakrabarti, Yixiao Li, Mengdi Wang, Tuo Zhao, Lin F. Yang
Abstract: Large language models (LLMs) exhibit remarkable performance across various natural language processing tasks but suffer from immense computational and memory demands, limiting their deployment in resource-constrained environments. To address this challenge, we propose NoWag (Normalized Weight and Activation Guided Compression), a unified framework for one-shot shape preserving compression algorithms. We apply NoWag to compress Llama-2 (7B, 13B, 70B) and Llama-3 (8B, 70B) models using two popular shape-preserving techniques: vector quantization (NoWag-VQ) and unstructured/semi-structured pruning (NoWag-P). Our results show that NoWag-VQ significantly outperforms state-of-the-art one-shot vector quantization methods, while NoWag-P performs competitively against leading pruning techniques. These findings highlight underlying commonalities between these compression paradigms and suggest promising directions for future research. Our code is available at https://github.com/LawrenceRLiu/NoWag
Authors: Weidi Luo, Tianyu Lu, Qiming Zhang, Xiaogeng Liu, Bin Hu, Yue Zhao, Jieyu Zhao, Song Gao, Patrick McDaniel, Zhen Xiang, Chaowei Xiao
Abstract: Recent advances in multi-modal large reasoning models (MLRMs) have shown significant ability to interpret complex visual content. While these models enable impressive reasoning capabilities, they also introduce novel and underexplored privacy risks. In this paper, we identify a novel category of privacy leakage in MLRMs: Adversaries can infer sensitive geolocation information, such as a user's home address or neighborhood, from user-generated images, including selfies captured in private settings. To formalize and evaluate these risks, we propose a three-level visual privacy risk framework that categorizes image content based on contextual sensitivity and potential for location inference. We further introduce DoxBench, a curated dataset of 500 real-world images reflecting diverse privacy scenarios. Our evaluation across 11 advanced MLRMs and MLLMs demonstrates that these models consistently outperform non-expert humans in geolocation inference and can effectively leak location-related private information. This significantly lowers the barrier for adversaries to obtain users' sensitive geolocation information. We further analyze and identify two primary factors contributing to this vulnerability: (1) MLRMs exhibit strong reasoning capabilities by leveraging visual clues in combination with their internal world knowledge; and (2) MLRMs frequently rely on privacy-related visual clues for inference without any built-in mechanisms to suppress or avoid such usage. To better understand and demonstrate real-world attack feasibility, we propose GeoMiner, a collaborative attack framework that decomposes the prediction process into two stages: clue extraction and reasoning to improve geolocation performance while introducing a novel attack perspective. Our findings highlight the urgent need to reassess inference-time privacy risks in MLRMs to better protect users' sensitive information.
Authors: Yu Cui, Feng Liu, Jiawei Chen, Xingyu Lou, Changwang Zhang, Jun Wang, Yuegang Sun, Xiaohu Yang, Can Wang
Abstract: Click-through rate (CTR) prediction is a fundamental task in modern recommender systems. In recent years, the integration of large language models (LLMs) has been shown to effectively enhance the performance of traditional CTR methods. However, existing LLM-enhanced methods often require extensive processing of detailed textual descriptions for large-scale instances or user/item entities, leading to substantial computational overhead. To address this challenge, this work introduces LLaCTR, a novel and lightweight LLM-enhanced CTR method that employs a field-level enhancement paradigm. Specifically, LLaCTR first utilizes LLMs to distill crucial and lightweight semantic knowledge from small-scale feature fields through self-supervised field-feature fine-tuning. Subsequently, it leverages this field-level semantic knowledge to enhance both feature representation and feature interactions. In our experiments, we integrate LLaCTR with six representative CTR models across four datasets, demonstrating its superior performance in terms of both effectiveness and efficiency compared to existing LLM-enhanced methods. Our code is available at https://github.com/istarryn/LLaCTR.
Authors: Jiaxin Liu, Jia Wang, Saihui Hou, Min Ren, Huijia Wu, Long Ma, Renwang Pei, Zhaofeng He
Abstract: In recent years, the explosive advancement of deepfake technology has posed a critical and escalating threat to public security: diffusion-based digital human generation. Unlike traditional face manipulation methods, such models can generate highly realistic videos with consistency via multimodal control signals. Their flexibility and covertness pose severe challenges to existing detection strategies. To bridge this gap, we introduce DigiFakeAV, the new large-scale multimodal digital human forgery dataset based on diffusion models. Leveraging five of the latest digital human generation methods and a voice cloning method, we systematically construct a dataset comprising 60,000 videos (8.4 million frames), covering multiple nationalities, skin tones, genders, and real-world scenarios, significantly enhancing data diversity and realism. User studies demonstrate that the misrecognition rate by participants for DigiFakeAV reaches as high as 68%. Moreover, the substantial performance degradation of existing detection models on our dataset further highlights its challenges. To address this problem, we propose DigiShield, an effective detection baseline based on spatiotemporal and cross-modal fusion. By jointly modeling the 3D spatiotemporal features of videos and the semantic-acoustic features of audio, DigiShield achieves state-of-the-art (SOTA) performance on the DigiFakeAV and shows strong generalization on other datasets.
Authors: Zhaolin Li, Jan Niehues
Abstract: With approximately 7,000 languages spoken worldwide, current large language models (LLMs) support only a small subset. Prior research indicates LLMs can learn new languages for certain tasks without supervised data. We extend this investigation to speech recognition, investigating whether LLMs can learn unseen, low-resource languages through in-context learning (ICL). With experiments on four diverse endangered languages that LLMs have not been trained on, we find that providing more relevant text samples enhances performance in both language modelling and Automatic Speech Recognition (ASR) tasks. Furthermore, we show that the probability-based approach outperforms the traditional instruction-based approach in language learning. Lastly, we show ICL enables LLMs to achieve ASR performance that is comparable to or even surpasses dedicated language models trained specifically for these languages, while preserving the original capabilities of the LLMs. Our code is publicly available.
Authors: Matthew Lisondra, Beno Benhabib, Goldie Nejat
Abstract: Rapid advancements in foundation models, including Large Language Models, Vision-Language Models, Multimodal Large Language Models, and Vision-Language-Action Models, have opened new avenues for embodied AI in mobile service robotics. By combining foundation models with the principles of embodied AI, where intelligent systems perceive, reason, and act through physical interaction, mobile service robots can achieve more flexible understanding, adaptive behavior, and robust task execution in dynamic real-world environments. Despite this progress, embodied AI for mobile service robots continues to face fundamental challenges related to the translation of natural language instructions into executable robot actions, multimodal perception in human-centered environments, uncertainty estimation for safe decision-making, and computational constraints for real-time onboard deployment. In this paper, we present the first systematic review focused specifically on the integration of foundation models in mobile service robotics. We analyze how recent advances in foundation models address these core challenges through language-conditioned control, multimodal sensor fusion, uncertainty-aware reasoning, and efficient model scaling. We further examine real-world applications in domestic assistance, healthcare, and service automation, highlighting how foundation models enable context-aware, socially responsive, and generalizable robot behaviors. Beyond technical considerations, we discuss ethical, societal, and human-interaction implications associated with deploying foundation model-enabled service robots in human environments. Finally, we outline future research directions emphasizing reliability and lifelong adaptation, privacy-aware and resource-constrained deployment, and governance and human-in-the-loop frameworks required for safe, scalable, and trustworthy mobile service robotics.
Authors: Peiling Jiang, Haijun Xia
Abstract: Web-based activities span multiple webpages. However, conventional browsers with stacks of tabs cannot support operating and synthesizing large volumes of information across pages. While recent AI systems enable fully automated web browsing and information synthesis, they often diminish user agency and hinder contextual understanding. We explore how AI could instead augment user interactions with content across webpages and mitigate cognitive and manual efforts. Through literature on information tasks and web browsing challenges, and an iterative design process, we present novel interactions with our prototype web browser, Orca. Leveraging AI, Orca supports user-driven exploration, operation, organization, and synthesis of web content at scale. To enable browsing at scale, webpages are treated as malleable materials that humans and AI can collaboratively manipulate and compose into a malleable, dynamic, and browser-level workspace. Our evaluation revealed an increased "appetite" for information foraging, enhanced control, and more flexible sensemaking across a broader web information landscape.
Authors: Qiang Chen, Tianyang Han, Jin Li, Ye Luo, Zigan Wang, Yuxiao Wu, Xiaowei Zhang, Tuo Zhou
Abstract: Can AI effectively perform complex econometric analysis traditionally requiring human expertise? This paper evaluates AI agents' capability to master econometrics, focusing on empirical analysis performance. We develop ``MetricsAI'', an Econometrics AI Agent built on the open-source MetaGPT framework. This agent exhibits outstanding performance in: (1) planning econometric tasks strategically, (2) generating and executing code, (3) employing error-based reflection for improved robustness, and (4) allowing iterative refinement through multi-round conversations. We construct two datasets from academic coursework materials and published research papers to evaluate performance against real-world challenges. Comparative testing shows our domain-specialized AI agent significantly outperforms both benchmark large language models (LLMs) and general-purpose AI agents. This work establishes a testbed for exploring AI's impact on social science research and enables cost-effective integration of domain expertise, making advanced econometric methods accessible to users with minimal coding skills. Furthermore, our AI agent enhances research reproducibility and offers promising pedagogical applications for econometrics teaching.
Authors: Shuang Chen, Yue Guo, Zhaochen Su, Yafu Li, Yulun Wu, Jiacheng Chen, Jiayu Chen, Weijie Wang, Xiaoye Qu, Yu Cheng
Abstract: Inspired by the remarkable reasoning capabilities of Deepseek-R1 in complex textual tasks, many works attempt to incentivize similar capabilities in Multimodal Large Language Models (MLLMs) by directly applying reinforcement learning (RL). However, they still struggle to activate complex reasoning. In this paper, rather than examining multimodal RL in isolation, we delve into current training pipelines and identify three crucial phenomena: 1) Effective cold start initialization is critical for enhancing MLLM reasoning. Intriguingly, we find that initializing with carefully selected text data alone can lead to performance surpassing many recent multimodal reasoning models, even before multimodal RL. 2) Standard GRPO applied to multimodal RL suffers from gradient stagnation, which degrades training stability and performance. 3) Subsequent text-only RL training, following the multimodal RL phase, further enhances multimodal reasoning. This staged training approach effectively balances perceptual grounding and cognitive reasoning development. By incorporating the above insights and addressing multimodal RL issues, we introduce ReVisual-R1, achieving a new state-of-the-art among open-source 7B MLLMs on challenging benchmarks including MathVerse, MathVision, WeMath, LogicVista, DynaMath, and challenging AIME2024 and AIME2025.
Authors: Hongzheng Chen, Yingheng Wang, Yaohui Cai, Hins Hu, Jiajie Li, Shirley Huang, Chenhui Deng, Rongjian Liang, Shufeng Kong, Haoxing Ren, Samitha Samaranayake, Carla P. Gomes, Zhiru Zhang
Abstract: While Large Language Models (LLMs) have demonstrated significant advancements in reasoning and agent-based problem-solving, current evaluation methodologies fail to adequately assess their capabilities: existing benchmarks either rely on closed-ended questions prone to saturation and memorization, or subjective comparisons that lack consistency and rigor. In this work, we introduce HeuriGym, an agentic framework designed for evaluating heuristic algorithms generated by LLMs for combinatorial optimization problems, characterized by clearly defined objectives and expansive solution spaces. HeuriGym empowers LLMs to propose heuristics, receive evaluative feedback via code execution, and iteratively refine their solutions. We evaluate nine state-of-the-art models on nine problems across domains such as computer systems, logistics, and biology, exposing persistent limitations in tool use, planning, and adaptive reasoning. To quantify performance, we propose the Quality-Yield Index (QYI), a metric that captures both solution pass rate and quality. Even top models like GPT-o4-mini-high and Gemini-2.5-Pro attain QYI scores of only 0.6, well below the expert baseline of 1. Our open-source benchmark aims to guide the development of LLMs toward more effective and realistic problem-solving in scientific and engineering domains.
Authors: Yang Zhang, Amr Mohamed, Hadi Abdine, Guokan Shang, Michalis Vazirgiannis
Abstract: Curriculum learning-organizing training data from easy to hard-has improved efficiency across machine learning domains, yet remains underexplored for language model pretraining. We present the first systematic investigation of curriculum learning in LLM pretraining, with over 200 models trained on up to 100B tokens across three strategies: vanilla curriculum learning, pacing-based sampling, and interleaved curricula, guided by six difficulty metrics spanning linguistic and information-theoretic properties. We evaluate performance on eight benchmarks under three realistic scenarios: limited data, unlimited data, and continual training. Our experiments show that curriculum learning consistently accelerates convergence in early and mid-training phases,reducing training steps by $18-45\%$ to reach baseline performance. When applied as a warmup strategy before standard random sampling, curriculum learning yields sustained improvements up to $3.5\%$. We identify compression ratio, lexical diversity (MTLD), and readability (Flesch Reading Ease) as the most effective difficulty signals. Our findings demonstrate that data ordering-orthogonal to existing data selection methods-provides a practical mechanism for more efficient LLM pretraining.
Authors: Bo-Cheng Chiu, Jen-Jee Chen, Yu-Chee Tseng, Feng-Chi Chen, An-Zi Yen
Abstract: Large Language Models (LLMs) have recently been extended to the video domain, enabling sophisticated video-language understanding. However, existing Video LLMs often exhibit limitations in fine-grained temporal reasoning, restricting their ability to precisely attribute responses to specific video moments, especially under constrained supervision. We introduce DaMO, a data-efficient Video LLM explicitly designed for accurate temporal reasoning and multimodal understanding. At its core, the proposed Temporal-aware Fuseformer employs a hierarchical dual-stream architecture that progressively captures temporal dynamics within each modality and effectively fuses complementary visual and audio information. To further enhance computational efficiency, DaMO integrates a global residual that reduces spatial redundancy while preserving essential semantic details. We train DaMO via a structured four-stage progressive training paradigm, incrementally equipping the model with multimodal alignment, semantic grounding, and temporal reasoning capabilities. This work also contributes multiple datasets augmented from existing ones with LLM-generated temporally grounded QA pairs for tasks requiring temporal supervision. Comprehensive experiments on temporal grounding and video QA benchmarks demonstrate that DaMO consistently surpasses prior methods, particularly in tasks demanding precise temporal alignment and reasoning. Our work establishes a promising direction for data-efficient video-language modeling.
Authors: Hao Shi, Xi Li, Fangfang Xie
Abstract: Black-box coevolution in mixed-motive games is often undermined by opponent-drift non-stationarity and noisy rollouts, which distort progress signals and can induce cycling, Red-Queen dynamics, and detachment. We propose the \emph{Marker Gene Method} (MGM), a curriculum-inspired governance mechanism that stabilizes selection by anchoring evaluation to cross-generational marker individuals, together with DWAM and conservative marker-update rules to reduce spurious updates. We also introduce NGD-Div, which adapts the key update threshold using a divergence proxy and natural-gradient optimization. We provide theoretical analysis in strictly competitive settings and evaluate MGM integrated with evolution strategies (MGM-E-NES) on coordination games and a resource-depletion Markov game. MGM-E-NES reliably recovers target coordination in Stag Hunt and Battle of the Sexes, achieving final cooperation probabilities close to $(1,1)$ (e.g., $0.991\pm0.01/1.00\pm0.00$ and $0.97\pm0.00/0.97\pm0.00$ for the two players). In the Markov resource game, it maintains high and stable state-conditioned cooperation across 30 seeds, with final cooperation of $\approx 0.954/0.980/0.916$ in \textsc{Rich}/\textsc{Poor}/\textsc{Collapsed} (both players; small standard deviations), indicating welfare-aligned and state-dependent behavior. Overall, MGM-E-NES transfers across tasks with minimal hyperparameter changes and yields consistently stable training dynamics, showing that top-level governance can substantially improve the robustness of black-box coevolution in dynamic environments.
Authors: Madhav Kotecha
Abstract: Online dating platforms have fundamentally transformed the formation of romantic relationships, with millions of users worldwide relying on algorithmic matching systems to find compatible partners. However, current recommendation systems in dating applications suffer from significant algorithmic deficiencies, including but not limited to popularity bias, filter bubble effects, and inadequate reciprocity modeling that limit effectiveness and introduce harmful biases. This research integrates foundational work with recent empirical findings to deliver a detailed analysis of dating app recommendation systems, highlighting key issues and suggesting research-backed solutions. Through analysis of reciprocal recommendation frameworks, fairness evaluation metrics, and industry implementations, we demonstrate that current systems achieve modest performance with collaborative filtering reaching 25.1\% while reciprocal methods achieve 28.7\%. Our proposed mathematical framework addresses these limitations through enhanced similarity measures, multi-objective optimization, and fairness-aware algorithms that maintain competitive accuracy while improving demographic representation to reduce algorithmic bias.
Authors: Hao Wang, Limeng Qiao, Zequn Jie, Zhijian Huang, Chengjian Feng, Qingfang Zheng, Lin Ma, Xiangyuan Lan, Xiaodan Liang
Abstract: Large Language Models (LLMs) demonstrate strong capabilities in broad knowledge representation, yet they are inherently deficient in pixel-level perceptual understanding. Although the Segment Anything Model (SAM) represents a significant advancement in visual-prompt-driven image segmentation, it exhibits notable limitations in multi-mask prediction and category-specific segmentation tasks, and it cannot integrate all segmentation tasks within a unified model architecture. To address these limitations, we present X-SAM, a streamlined Multimodal Large Language Model (MLLM) framework that extends the segmentation paradigm from \textit{segment anything} to \textit{any segmentation}. Specifically, we introduce a novel unified framework that enables more advanced pixel-level perceptual comprehension for MLLMs. Furthermore, we propose a new segmentation task, termed Visual GrounDed (VGD) segmentation, which segments all instance objects with interactive visual prompts and empowers MLLMs with visual grounded, pixel-wise interpretative capabilities. To enable effective training on diverse data sources, we present a unified training strategy that supports co-training across multiple datasets. Experimental results demonstrate that X-SAM achieves state-of-the-art performance on a wide range of image segmentation benchmarks, highlighting its efficiency for multimodal, pixel-level visual understanding. Code is available at https://github.com/wanghao9610/X-SAM.
Authors: Yu Liu, Yanbing Liu, Fangfang Yuan, Cong Cao, Youbang Sun, Kun Peng, WeiZhuo Chen, Jianjun Li, Zhiyuan Ma
Abstract: Recent advances in large language models (LLMs) and dense retrievers have driven significant progress in retrieval-augmented generation (RAG). However, existing approaches face significant challenges in complex reasoning-oriented multi-hop retrieval tasks: 1) Ineffective reasoning-oriented planning: Prior methods struggle to generate robust multi-step plans for complex queries, as rule-based decomposers perform poorly on out-of-template questions. 2) Suboptimal reasoning-driven retrieval: Related methods employ limited query reformulation, leading to iterative retrieval loops that often fail to locate golden documents. 3) Insufficient reasoning-guided filtering: Prevailing methods lack the fine-grained reasoning to effectively filter salient information from noisy results, hindering utilization of retrieved knowledge. Fundamentally, these limitations all stem from the weak coupling between retrieval and reasoning in current RAG architectures. We introduce the Orchestrated Planner-Executor Reasoning Architecture (OPERA), a novel reasoning-driven retrieval framework. OPERA's Goal Planning Module (GPM) decomposes questions into sub-goals, which are executed by a Reason-Execute Module (REM) with specialized components for precise reasoning and effective retrieval. To train OPERA, we propose Multi-Agents Progressive Group Relative Policy Optimization (MAPGRPO), a novel variant of GRPO. Experiments on complex multi-hop benchmarks show OPERA's superior performance, validating both the MAPGRPO method and OPERA's design.
Authors: Junjie Mu, Zonghao Ying, Zhekui Fan, Zonglei Jing, Yaoyuan Zhang, Zhengmin Yu, Wenxin Zhang, Quanchen Zou, Xiangzheng Zhang
Abstract: Jailbreak attacks on Large Language Models (LLMs) have demonstrated various successful methods whereby attackers manipulate models into generating harmful responses that they are designed to avoid. Among these, Greedy Coordinate Gradient (GCG) has emerged as a general and effective approach that optimizes the tokens in a suffix to generate jailbreakable prompts. While several improved variants of GCG have been proposed, they all rely on fixed-length suffixes. However, the potential redundancy within these suffixes remains unexplored. In this work, we propose Mask-GCG, a plug-and-play method that employs learnable token masking to identify impactful tokens within the suffix. Our approach increases the update probability for tokens at high-impact positions while pruning those at low-impact positions. This pruning not only reduces redundancy but also decreases the size of the gradient space, thereby lowering computational overhead and shortening the time required to achieve successful attacks compared to GCG. We evaluate Mask-GCG by applying it to the original GCG and several improved variants. Experimental results show that most tokens in the suffix contribute significantly to attack success, and pruning a minority of low-impact tokens does not affect the loss values or compromise the attack success rate (ASR), thereby revealing token redundancy in LLM prompts. Our findings provide insights for developing efficient and interpretable LLMs from the perspective of jailbreak attacks.
Authors: Yuecheng Liu, Dafeng Chi, Shiguang Wu, Zhanguang Zhang, Yuzheng Zhuang, Bowen Yang, He Zhu, Lingfeng Zhang, Pengwei Xie, David Gamaliel Arcos Bravo, Yingxue Zhang, Jianye Hao, Xingyue Quan
Abstract: Recent advances in multimodal large language models (MLLMs) have opened new opportunities for embodied intelligence, enabling multimodal understanding, reasoning, and interaction, as well as continuous spatial decision-making. Nevertheless, current MLLM-based embodied systems face two critical limitations. First, Geometric Adaptability Gap: models trained solely on 2D inputs or with hard-coded 3D geometry injection suffer from either insufficient spatial information or restricted 2D generalization, leading to poor adaptability across tasks with diverse spatial demands. Second, Embodiment Constraint Gap: prior work often neglects the physical constraints and capacities of real robots, resulting in task plans that are theoretically valid but practically infeasible. To address these gaps, we introduce OmniEVA -- an embodied versatile planner that enables advanced embodied reasoning and task planning through two pivotal innovations: (1) a Task-Adaptive 3D Grounding mechanism, which introduces a gated router to perform explicit selective regulation of 3D fusion based on contextual requirements, enabling context-aware 3D grounding for diverse embodied tasks. (2) an Embodiment-Aware Reasoning framework that jointly incorporates task goals and embodiment constraints into the reasoning loop, resulting in planning decisions that are both goal-directed and executable. Extensive experimental results demonstrate that OmniEVA not only achieves state-of-the-art general embodied reasoning performance, but also exhibits a strong ability across a wide range of downstream scenarios. Evaluations of a suite of proposed embodied benchmarks, including both primitive and composite tasks, confirm its robust and versatile planning capabilities. Project page: https://omnieva.github.io
Authors: Tao Long, Xuanming Zhang, Sitong Wang, Zhou Yu, Lydia B Chilton
Abstract: Agentic workflows promise efficiency, but adoption hinges on whether people can align systems that act on their behalf with their goals, values, and situational expectations. We present DoubleAgents, an agentic planning tool that embeds transparency and control through user intervention, value-reflecting policies, rich state visualizations, and uncertainty flagging for human coordination tasks. A built-in respondent simulation generates realistic scenarios, allowing users to rehearse and refine policies and calibrate their use of agentic behavior before live deployment. We evaluate DoubleAgents in a two-day lab study (n = 10), three deployment studies, and a technical evaluation. Results show that participants initially hesitated to delegate but used simulation to probe system behavior and adjust policies, gradually increasing delegation as agent actions became better aligned with their intentions and context. Deployment results demonstrate DoubleAgents' real-world relevance and usefulness, showing that simulation helps users effectively manage real-world tasks with higher complexity and uncertainty. We contribute interactive simulation as a practical pathway for users to iteratively align and calibrate agentic systems.
Authors: Graham Clyne, Guillaume Couairon, Guillaume Gastineau, Claire Monteleoni, Anastase Charantonis
Abstract: Climate projections have uncertainties related to components of the climate system and their interactions. A typical approach to quantifying these uncertainties is to use climate models to create ensembles of repeated simulations under different initial conditions. Due to the complexity of these simulations, generating such ensembles of projections is computationally expensive. In this work, we present ArchesClimate, a deep learning-based climate model emulator that aims to reduce this cost. ArchesClimate is trained on decadal hindcasts of the IPSL-CM6A-LR climate model at a spatial resolution of approximately 2.5x1.25 degrees. We train a flow matching model following ArchesWeatherGen, which we adapt to predict near-term climate. Once trained, the model generates states at a one-month lead time and can be used to auto-regressively emulate climate model simulations of any length. We show that for up to 10 years, these generations are stable and physically consistent. We also show that for several important climate variables, ArchesClimate generates simulations that are interchangeable with the IPSL model. This work suggests that climate model emulators could significantly reduce the cost of climate model simulations.
Authors: Hyunjong Ok, Suho Yoo, Hyeonjun Kim, Jaeho Lee
Abstract: Even without directly hearing sounds, humans can effortlessly reason about auditory properties, such as pitch, loudness, or sound-source associations, drawing on auditory commonsense. In contrast, language models often lack this capability, limiting their effectiveness in multimodal interactions. As an initial step to address this gap, we present AuditoryBench++, a comprehensive benchmark for evaluating auditory knowledge and reasoning in text-only settings. The benchmark encompasses tasks that range from basic auditory comparisons to contextually grounded reasoning, enabling fine-grained analysis of how models process and integrate auditory concepts. In addition, we introduce AIR-CoT, a novel auditory imagination reasoning method that generates and integrates auditory information during inference through span detection with special tokens and knowledge injection. Extensive experiments with recent LLMs and Multimodal LLMs demonstrate that AIR-CoT generally outperforms both the off-the-shelf models and those augmented with auditory knowledge. The project page is available at https://auditorybenchpp.github.io.
Authors: Xinyu He, Xianghui Cao
Abstract: Post-training fundamentally alters the behavior of large language models (LLMs), yet its impact on the internal parameter space remains poorly understood. In this work, we conduct a systematic singular value decomposition (SVD) analysis of principal linear layers in pretrained LLMs, focusing on two widely adopted post-training methods: instruction tuning and long-chain-of-thought (Long-CoT) distillation. Our analysis reveals two unexpected and robust structural changes: (1) a near-uniform geometric scaling of singular values across layers; and (2) highly consistent orthogonal transformations are applied to the left and right singular vectors of each matrix. Based on these findings, We propose a simple yet effective framework to describe the coordinated dynamics of parameters in LLMs, which elucidates why post-training inherently relies on the foundational capabilities developed during pre-training. Further experiments demonstrate that singular value scaling underpins the temperature-controlled regulatory mechanisms of post-training, while the coordinated rotation of singular vectors encodes the essential semantic alignment. These results challenge the prevailing view of the parameter space in large models as a black box, uncovering the first clear regularities in how parameters evolve during training, and providing a new perspective for deeper investigation into model parameter changes.
Authors: Duc-Tuan Truong, Tianchi Liu, Junjie Li, Ruijie Tao, Kong Aik Lee, Eng Siong Chng
Abstract: In speech deepfake detection (SDD), data augmentation (DA) is commonly used to improve model generalization across varied speech conditions and spoofing attacks. However, during training, the backpropagated gradients from original and augmented inputs may misalign, which can result in conflicting parameter updates. These conflicts could hinder convergence and push the model toward suboptimal solutions, thereby reducing the benefits of DA. To investigate and address this issue, we design a dual-path data-augmented (DPDA) training framework with gradient alignment for SDD. In our framework, each training utterance is processed through two input paths: one using the original speech and the other with its augmented version. This design allows us to compare and align their backpropagated gradient directions to reduce optimization conflicts. Our analysis shows that approximately 25% of training iterations exhibit gradient conflicts between the original inputs and their augmented counterparts when using RawBoost augmentation. By resolving these conflicts with gradient alignment, our method accelerates convergence by reducing the number of training epochs and achieves up to an 18.69% relative reduction in Equal Error Rate on the In-the-Wild dataset compared to the baseline.
Authors: Geunhyeok Yu, Sunjae Jeong, Yoonyoung Choi, Jaeseung Kim, Hyoseok Hwang
Abstract: Vision Transformers are widely adopted as the backbone of vision foundation models, but they are known to produce high-norm artifacts that degrade representation quality. When knowledge distillation transfers these features to students, high-norm artifacts dominate the objective, so students overfit to artifacts and underweight informative signals, diminishing the gains from larger models. Prior work attempted to remove artifacts but encountered an inherent trade-off between artifact suppression and preserving informative signals from teachers. To address this, we introduce Singular Nullspace-Guided Energy Reallocation (SiNGER), a novel distillation framework that suppresses artifacts while preserving informative signals. The key idea is principled teacher feature refinement: during refinement, we leverage the nullspace-guided perturbation to preserve information while suppressing artifacts. Then, the refined teacher's features are distilled to a student. We implement this perturbation efficiently with a LoRA-based adapter that requires minimal structural modification. Extensive experiments show that \oursname consistently improves student models, achieving state-of-the-art performance in multiple downstream tasks and producing clearer and more interpretable representations.
Authors: Hakaze Cho, Haolin Yang, Gouki Minegishi, Naoya Inoue
Abstract: In-context Learning (ICL) is an emerging few-shot learning paradigm based on modern Language Models (LMs), yet its inner mechanism remains unclear. In this paper, we investigate the mechanism through a novel perspective of information removal. Specifically, we demonstrate that in the zero-shot scenario, LMs encode queries into non-selective representations in hidden states containing information for all possible tasks, leading to arbitrary outputs without focusing on the intended task, resulting in near-zero accuracy. Meanwhile, we find that selectively removing specific information from hidden states by a low-rank filter effectively steers LMs toward the intended task. Building on these findings, by measuring the hidden states on carefully designed metrics, we observe that few-shot ICL effectively simulates such task-oriented information removal processes, selectively removing the redundant information from entangled non-selective representations, and improving the output based on the demonstrations, which constitutes a key mechanism underlying ICL. Moreover, we identify essential attention heads inducing the removal operation, termed Denoising Heads, which enables the ablation experiments blocking the information removal operation from the inference, where the ICL accuracy significantly degrades, especially when the correct label is absent from the few-shot demonstrations, confirming both the critical role of the information removal mechanism and denoising heads.
Authors: Miao Jing, Mengting Jia, Junling Lin, Zhongxia Shen, Huan Gao, Mingkun Xu, Shangyang Li
Abstract: Recent advances in vision-language models (VLMs) have achieved remarkable performance on standard medical benchmarks, yet their true clinical reasoning ability remains unclear. Existing datasets predominantly emphasize classification accuracy, creating an evaluation illusion in which models appear proficient while still failing at high-stakes diagnostic reasoning. We introduce Neural-MedBench, a compact yet reasoning-intensive benchmark specifically designed to probe the limits of multimodal clinical reasoning in neurology. Neural-MedBench integrates multi-sequence MRI scans, structured electronic health records, and clinical notes, and encompasses three core task families: differential diagnosis, lesion recognition, and rationale generation. To ensure reliable evaluation, we develop a hybrid scoring pipeline that combines LLM-based graders, clinician validation, and semantic similarity metrics. Through systematic evaluation of state-of-the-art VLMs, including GPT-4o, Claude-4, and MedGemma, we observe a sharp performance drop compared to conventional datasets. Error analysis shows that reasoning failures, rather than perceptual errors, dominate model shortcomings. Our findings highlight the necessity of a Two-Axis Evaluation Framework: breadth-oriented large datasets for statistical generalization, and depth-oriented, compact benchmarks such as Neural-MedBench for reasoning fidelity. We release Neural-MedBench at https://neuromedbench.github.io/ as an open and extensible diagnostic testbed, which guides the expansion of future benchmarks and enables rigorous yet cost-effective assessment of clinically trustworthy AI.
Authors: Yaorui Shi, Yuxin Chen, Siyuan Wang, Sihang Li, Hengxing Cai, Qi Gu, Xiang Wang, An Zhang
Abstract: Large language models face challenges in long-context question answering, where key evidence of a query may be dispersed across millions of tokens. Existing works equip large language models with a memory buffer that is dynamically updated via a linear document scan, also known as the "memorize while reading" methods. While this approach scales efficiently, it suffers from pruning of latent evidence, information loss through overwriting, and sparse reinforcement learning signals. To tackle these challenges, we present ReMemR1, which integrates the mechanism of memory retrieval into the memory update process, enabling the agent to selectively callback historical memories for non-linear reasoning. To further strengthen training, we propose a multi-level reward design, which combines final-answer rewards with dense, step-level signals that guide effective memory use. Together, these contributions mitigate information degradation, improve supervision, and support complex multi-hop reasoning. Extensive experiments demonstrate that ReMemR1 significantly outperforms state-of-the-art baselines on long-context question answering while incurring negligible computational overhead, validating its ability to trade marginal cost for robust long-context reasoning.
Authors: Michael Drolet, Firas Al-Hafez, Aditya Bhatt, Jan Peters, Oleg Arenz
Abstract: Discrete latent bottlenecks in variational autoencoders (VAEs) offer high bit efficiency and can be modeled with autoregressive discrete distributions, enabling parameter-efficient multimodal search with transformers. However, discrete random variables do not allow for exact differentiable parameterization; therefore, discrete VAEs typically rely on approximations, such as Gumbel-Softmax reparameterization or straight-through gradient estimates, or employ high-variance gradient-free methods such as REINFORCE that have had limited success on high-dimensional tasks such as image reconstruction. Inspired by popular techniques in policy search, we propose a training framework for discrete VAEs that leverages the natural gradient of a non-parametric encoder to update the parametric encoder without requiring reparameterization. Our method, combined with automatic step size adaptation and a transformer-based encoder, scales to challenging datasets such as ImageNet and outperforms both approximate reparameterization methods and quantization-based discrete autoencoders in reconstructing high-dimensional data from compact latent spaces.
Authors: David Anugraha, Shou-Yi Hung, Zilu Tang, Annie En-Shiun Lee, Derry Tanti Wijaya, Genta Indra Winata
Abstract: Evaluation using Large Language Model (LLM) judges has been widely adopted in English and shown to be effective for automatic evaluation. However, their performance does not generalize well to non-English settings, and it remains unclear what constitutes effective multilingual training for such judges. In this paper, we introduce mR3, a massively multilingual, rubric-agnostic reward reasoning model trained on 72 languages, achieving the broadest language coverage in reward modeling to date. We present a comprehensive study of data and curriculum selection for training to identify effective strategies and data sources for building high-quality reward models, including support for reasoning in the target language. Our approach attains state-of-the-art performance on multilingual reward model benchmarks, surpassing much larger models (i.e., GPT-OSS-120B) while being up to 9x smaller, and its effectiveness is further confirmed through extensive ablation studies. Finally, we demonstrate the effectiveness of mR3 in off-policy preference optimization and validate the quality of its reasoning traces and rubric-based evaluations through human studies with 20 annotators across 12 languages, where mR3 models' reasoning is preferred, including for extremely low-resource languages that are entirely unseen during training. Our models, data, and code are available as open source at https://github.com/rubricreward/mr3.
Authors: Silvia Sapora, Devon Hjelm, Alexander Toshev, Omar Attia, Bogdan Mazoure
Abstract: Inverse Reinforcement Learning aims to recover reward models from expert demonstrations, but traditional methods yield black-box models that are difficult to interpret and debug. In this work, we introduce GRACE (Generating Rewards As CodE), a method for using Large Language Models within an evolutionary search to reverse-engineer an interpretable, code-based reward function directly from expert trajectories. The resulting reward function is executable code that can be inspected and verified. We empirically validate GRACE on the MuJoCo, BabyAI and AndroidWorld benchmarks, where it efficiently learns highly accurate rewards, even in complex, multi-task settings. Further, we demonstrate that the resulting reward leads to strong policies, compared to both competitive Imitation Learning and online RL approaches with ground-truth rewards. Finally, we show that GRACE is able to build complex reward APIs in multi-task setups.
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: Florent Delgrange, Raphael Avalos, Willem R\"opke
Abstract: Safe policy improvement (SPI) offers theoretical control over policy updates, yet existing guarantees largely concern offline, tabular reinforcement learning (RL). We study SPI in general online settings, when combined with world model and representation learning. We develop a theoretical framework showing that restricting policy updates to a well-defined neighborhood of the current policy ensures monotonic improvement and convergence. This analysis links transition and reward prediction losses to representation quality, yielding online, "deep" analogues of classical SPI theorems from the offline RL literature. Building on these results, we introduce DeepSPI, a principled on-policy algorithm that couples local transition and reward losses with regularised policy updates. On the ALE-57 benchmark, DeepSPI matches or exceeds strong baselines, including PPO and DeepMDPs, while retaining theoretical guarantees.
Authors: Chao Chen, Zhixin Ma, Yongqi Li, Yupeng Hu, Yinwei Wei, Wenjie Li, Liqiang Nie
Abstract: Multimodal reasoning aims to enhance the capabilities of MLLMs by incorporating intermediate reasoning steps before reaching the final answer. It has evolved from text-only reasoning to the integration of visual information, enabling the thought process to be conveyed through both images and text. Despite its effectiveness, current multimodal reasoning methods depend on explicit reasoning steps that require labor-intensive vision-text annotations and inherently introduce significant inference latency. To address these issues, we introduce multimodal latent reasoning with the advantages of multimodal representation, reduced annotation, and inference efficiency. To facilitate it, we propose Interleaved Vision-Text Latent Reasoning (IVT-LR), which injects both visual and textual information in the reasoning process within the latent space. Specifically, IVT-LR represents each reasoning step by combining two implicit parts: latent text (the hidden states from the previous step) and latent vision (a set of selected image embeddings). We further introduce a progressive multi-stage training strategy to enable MLLMs to perform the above multimodal latent reasoning steps. Experiments on M$^3$CoT and ScienceQA demonstrate that our IVT-LR method achieves an average performance increase of 5.45\% in accuracy, while simultaneously achieving a speed increase of over 5 times compared to existing approaches.
Authors: Shuangshuang Ying, Yunwen Li, Xingwei Qu, Xin Li, Sheng Jin, Minghao Liu, Zhoufutu Wen, Xeron Du, Tianyu Zheng, Yichi Zhang, Letian Ni, Yuyang Cheng, Zhenzhu Yang, Qiguang Chen, Jingzhe Ding, Shengda Long, Wangchunshu Zhou, Jiazhan Feng, Wanjun Zhong, Libo Qin, Ge Zhang, Wenhao Huang, Wanxiang Che, Chenghua Lin
Abstract: Current preference learning methods achieve high accuracy on standard benchmarks but exhibit significant performance degradation when objective quality signals are removed. We introduce WritingPreferenceBench, a dataset of 1,800 human-annotated preference pairs (1,200 English, 600 Chinese) across 8 creative writing genres, where responses are matched for objective correctness, factual accuracy, and length. On this benchmark, sequence-based reward models--the standard architecture for RLHF--achieve only 52.7% mean accuracy, while zero-shot language model judges perform at 53.9%. In contrast, generative reward models that produce explicit reasoning chains achieve 81.8% accuracy. We observe high within-model variance across genres: individual models range from 18.2% to 81.8% accuracy across different writing categories, with standard deviations averaging 10.1%. This variance persists regardless of model scale, with 27B parameter models showing no consistent improvement over 8B variants. Our results suggest that current RLHF methods primarily learn to detect objective errors rather than capture subjective quality preferences (e.g., creativity, stylistic flair, and emotional resonance), and that successful preference modeling may require intermediate reasoning representations rather than direct classification.
Authors: Bernardo Torres, Manuel Moussallam, Gabriel Meseguer-Brocal
Abstract: Audio autoencoders learn useful, compressed audio representations, but their non-linear latent spaces prevent intuitive algebraic manipulation such as mixing or scaling. We introduce a simple training methodology to induce linearity in a high-compression Consistency Autoencoder (CAE) by using data augmentation, thereby inducing homogeneity (equivariance to scalar gain) and additivity (the decoder preserves addition) without altering the model's architecture or loss function. When trained with our method, the CAE exhibits linear behavior in both the encoder and decoder while preserving reconstruction fidelity. We test the practical utility of our learned space on music source composition and separation via simple latent arithmetic. This work presents a straightforward technique for constructing structured latent spaces, enabling more intuitive and efficient audio processing.
Authors: Rui Xing, Preslav Nakov, Timothy Baldwin, Jey Han Lau
Abstract: Fact-checking on major platforms, such as X, Meta, and TikTok, is shifting from expert-driven verification to a community-based setup, where users contribute explanatory notes to clarify why a post might be misleading. An important challenge here is determining whether an explanation is helpful for understanding real-world claims and the reasons why, which remains largely underexplored in prior research. In practice, most community notes remain unpublished due to slow community annotation, and the reasons for helpfulness lack clear definitions. To bridge these gaps, we introduce the task of predicting both the helpfulness of explanatory notes and the reason for this. We present COMMUNITYNOTES, a large-scale multilingual dataset of 104k posts with user-provided notes and helpfulness labels. We further propose a framework that automatically generates and improves reason definitions via automatic prompt optimization, and integrate them into prediction. Our experiments show that the optimized definitions can improve both helpfulness and reason prediction. Finally, we show that the helpfulness information is beneficial for existing fact-checking systems.
Authors: Dmytro Vitel, Anshuman Chhabra
Abstract: Identifying how training samples influence/impact Large Language Model (LLM) decision-making is essential for effectively interpreting model decisions and auditing large-scale datasets. Current training sample influence estimation methods (also known as influence functions) undertake this goal by utilizing information flow through the model via its first-order and higher-order gradient terms. However, owing to the large model sizes of today consisting of billions of parameters, these influence computations are often restricted to some subset of model layers to ensure computational feasibility. Prior seminal work by Yeh et al. (2022) in assessing which layers are best suited for computing language data influence concluded that the first (embedding) layers are the most informative for this purpose, using a hypothesis based on influence scores canceling out (i.e., the cancellation effect). In this work, we propose theoretical and empirical evidence demonstrating how the cancellation effect is unreliable, and that middle attention layers are better estimators for influence. Furthermore, we address the broader challenge of aggregating influence scores across layers, and showcase how alternatives to standard averaging (such as ranking and vote-based methods) can lead to significantly improved performance. Finally, we propose better methods for evaluating influence score efficacy in LLMs without undertaking model retraining, and propose a new metric known as the Noise Detection Rate (NDR) that exhibits strong predictive capability compared to the cancellation effect. Through extensive experiments across LLMs of varying types and scales, we concretely determine that the first (layers) are not necessarily better than the last (layers) for LLM influence estimation, contrasting with prior knowledge in the field.
Authors: Daojun Liang, Jing Chen, Xiao Wang, Yinglong Wang, Shuo Li
Abstract: Time-Series (TS) exhibits pronounced non-stationarity. Consequently, most forecasting methods display compromised robustness to concept drift, despite the prevalent application of instance normalization. We tackle this challenge by first analysing concept drift through a bias-variance lens and proving that weighted ensemble reduces variance without increasing bias. These insights motivate DeepBooTS, a novel end-to-end dual-stream residual-decreasing boosting method that progressively reconstructs the intrinsic signal. In our design, each block of a deep model becomes an ensemble of learners with an auxiliary output branch forming a highway to the final prediction. The block-wise outputs correct the residuals of previous blocks, leading to a learning-driven decomposition of both inputs and targets. This method enhances versatility and interpretability while substantially improving robustness to concept drift. Extensive experiments, including those on large-scale datasets, show that the proposed method outperforms existing methods by a large margin, yielding an average performance improvement of 15.8% across various datasets, establishing a new benchmark for TS forecasting.
Authors: Linfeng Dong, Yuchen Yang, Hao Wu, Wei Wang, Yuenan Hou, Zhihang Zhong, Xiao Sun
Abstract: We introduce RacketVision, a novel dataset and benchmark for advancing computer vision in sports analytics, covering table tennis, tennis, and badminton. The dataset is the first to provide large-scale, fine-grained annotations for racket pose alongside traditional ball positions, enabling research into complex human-object interactions. It is designed to tackle three interconnected tasks: fine-grained ball tracking, articulated racket pose estimation, and predictive ball trajectory forecasting. Our evaluation of established baselines reveals a critical insight for multi-modal fusion: while naively concatenating racket pose features degrades performance, a CrossAttention mechanism is essential to unlock their value, leading to trajectory prediction results that surpass strong unimodal baselines. RacketVision provides a versatile resource and a strong starting point for future research in dynamic object tracking, conditional motion forecasting, and multimodal analysis in sports. Project page at https://github.com/OrcustD/RacketVision
Authors: Jungyang Park, Suho Kang, Jaewoo Park, Jaehong Kim, Jaewoo Shin, Seonjoon Park, Youngjae Yu
Abstract: Knowledge Tracing (KT) aims to model student's knowledge state and predict future performance to enable personalized learning in Intelligent Tutoring Systems. However, traditional KT methods face fundamental limitations in explainability, as they rely solely on the response correctness, neglecting the rich information embedded in students' problem-solving processes. To address this gap, we propose Knowledge Tracing Leveraging Problem-Solving Process (KT-PSP), which incorporates students' problem-solving processes to capture the multidimensional aspects of mathematical proficiency. We also introduce KT-PSP-25, a new dataset specifically designed for the KT-PSP. Building on this, we present StatusKT, a KT framework that employs a teacher-student-teacher three-stage LLM pipeline to extract students' MP as intermediate signals. In this pipeline, the teacher LLM first extracts problem-specific proficiency indicators, then a student LLM generates responses based on the student's solution process, and a teacher LLM evaluates these responses to determine mastery of each indicator. The experimental results on KT-PSP-25 demonstrate that StatusKT improves the prediction performance of existing KT methods. Moreover, StatusKT provides interpretable explanations for its predictions by explicitly modeling students' mathematical proficiency.
Authors: Wei Chee Yew, Hailun Xu, Sanjay Saha, Xiaotian Fan, Hiok Hian Ong, David Yuchen Wang, Kanchan Sarkar, Zhenheng Yang, Danhui Guan
Abstract: Content moderation remains a critical yet challenging task for large-scale user-generated video platforms, especially in livestreaming environments where moderation must be timely, multimodal, and robust to evolving forms of unwanted content. We present a hybrid moderation framework deployed at production scale that combines supervised classification for known violations with reference-based similarity matching for novel or subtle cases. This hybrid design enables robust detection of both explicit violations and novel edge cases that evade traditional classifiers. Multimodal inputs (text, audio, visual) are processed through both pipelines, with a multimodal large language model (MLLM) distilling knowledge into each to boost accuracy while keeping inference lightweight. In production, the classification pipeline achieves 67% recall at 80% precision, and the similarity pipeline achieves 76% recall at 80% precision. Large-scale A/B tests show a 6-8% reduction in user views of unwanted livestreams}. These results demonstrate a scalable and adaptable approach to multimodal content governance, capable of addressing both explicit violations and emerging adversarial behaviors.
Authors: Adnan Munir, Muhammad Shahid Jabbar, Shujaat Khan
Abstract: Medical image segmentation plays a pivotal role in automated diagnostic and treatment planning systems. In this work, we present DAUNet, a novel lightweight UNet variant that integrates Deformable V2 Convolutions and Parameter-Free Attention (SimAM) to improve spatial adaptability and context-aware feature fusion without increasing model complexity. DAUNet's bottleneck employs dynamic deformable kernels to handle geometric variations, while the decoder and skip pathways are enhanced using SimAM attention modules for saliency-aware refinement. Extensive evaluations on two challenging datasets, FH-PS-AoP (fetal head and pubic symphysis ultrasound) and FUMPE (CT-based pulmonary embolism detection), demonstrate that DAUNet outperforms state-of-the-art models in Dice score, HD95, and ASD, while maintaining superior parameter efficiency. Ablation studies highlight the individual contributions of deformable convolutions and SimAM attention. DAUNet's robustness to missing context and low-contrast regions establishes its suitability for deployment in real-time and resource-constrained clinical environments.
Authors: David Samuel, Lucas Georges Gabriel Charpentier
Abstract: This paper combines autoregressive and masked-diffusion training objectives without any architectural modifications, resulting in flexible language models that outperform single-objective models. Autoregressive modeling has been a popular approach, partly because of its training efficiency; however, that comes at the cost of sensitivity to overfitting. On the other hand, masked-diffusion models are less efficient to train while being more resilient to overfitting. In this work, we demonstrate that dual-objective training achieves the best of both worlds. To derive the optimal balance between both objectives, we train and evaluate 50 language models under varying levels of data repetition. We show that it is optimal to combine both objectives under all evaluated settings and that the optimal balance is similar whether targeting autoregressive or masked-diffusion downstream performance.
Authors: Chenxiao Gao, Haotian Sun, Na Li, Dale Schuurmans, Bo Dai
Abstract: In real-world applications with large state and action spaces, reinforcement learning (RL) typically employs function approximations to represent core components like the policies, value functions, and dynamics models. Although powerful approximations such as neural networks offer great expressiveness, they often present theoretical ambiguities, suffer from optimization instability and exploration difficulty, and incur substantial computational costs in practice. In this paper, we introduce the perspective of spectral representations as a solution to address these difficulties in RL. Stemming from the spectral decomposition of the transition operator, this framework yields an effective abstraction of the system dynamics for subsequent policy optimization while also providing a clear theoretical characterization. We reveal how to construct spectral representations for transition operators that possess latent variable structures or energy-based structures, which implies different learning methods to extract spectral representations from data. Notably, each of these learning methods realizes an effective RL algorithm under this framework. We also provably extend this spectral view to partially observable MDPs. Finally, we validate these algorithms on over 20 challenging tasks from the DeepMind Control Suite, where they achieve performances comparable or superior to current state-of-the-art model-free and model-based baselines.
Authors: Yen-Chieh Huang, Pi-Cheng Hsiu, Rui Fang, Ming-Syan Chen
Abstract: Long-context LLM inference is bottlenecked by the quadratic attention complexity and linear KV cache growth. Prior approaches mitigate this via post-hoc selection or eviction but overlook the root inefficiency: indiscriminate writing to memory. In this paper, we formalize KV cache management as a causal system of three primitives: KV Admission, Selection, and Eviction. We instantiate KV Admission via Write-Gated KV (WG-KV), a lightweight mechanism that learns to predict token utility before cache entry. By filtering out low-utility states early to maintain a compact global cache alongside a sliding local cache, WG-KV reduces memory usage by 46-68% and delivers 3.03-3.70x prefill and 1.85-2.56x decode speedups on Llama and Qwen models, while maintaining compatibility with FlashAttention and Paged-KV systems. These results demonstrate that learning what to write is a principled and practical recipe for efficient long-context inference. Code is available at https://github.com/EMCLab-Sinica/WG-KV.
Authors: Yueru Yan, Tuc Nguyen, Bo Su, Melissa Lieffers, Thai Le
Abstract: While academic research typically treats Large Language Models (LLM) as generic text generators, they are distinct commercial products with unique interfaces and capabilities that fundamentally shape user behavior. Current datasets obscure this reality by collecting text-only data through uniform interfaces that fail to capture authentic chatbot usage. To address this limitation, we present ShareChat, a large-scale corpus of 142,808 conversations (660,293 turns) sourced directly from publicly shared URLs on ChatGPT, Perplexity, Grok, Gemini, and Claude. ShareChat distinguishes itself by preserving native platform affordances, such as citations and thinking traces, across a diverse collection covering 101 languages and the period from April 2023 to October 2025. Furthermore, ShareChat offers substantially longer context windows and greater interaction depth than prior datasets. To illustrate the dataset's breadth, we present three case studies: a completeness analysis of intent satisfaction, a citation study of model grounding, and a temporal analysis of engagement rhythms. This work provides the community with a vital and timely resource for understanding authentic user-LLM chatbot interactions in the wild. The dataset is publicly available via Hugging Face.
Authors: Jiayun Wu, Jiashuo Liu, Zhiyuan Zeng, Tianyang Zhan, Tianle Cai, Wenhao Huang
Abstract: LLM deployment in critical domains is currently impeded by persistent hallucinations--generating plausible but factually incorrect assertions. While scaling laws drove significant improvements in general capabilities, theoretical frameworks suggest hallucination is not merely stochastic error but a predictable statistical consequence of training objectives prioritizing mimicking data distribution over epistemic honesty. Standard RLVR paradigms, utilizing binary reward signals, inadvertently incentivize models as good test-takers rather than honest communicators, encouraging guessing whenever correctness probability exceeds zero. This paper presents an exhaustive investigation into behavioral calibration, which incentivizes models to stochastically admit uncertainty by abstaining when not confident, aligning model behavior with accuracy. Synthesizing recent advances, we propose and evaluate training interventions optimizing strictly proper scoring rules for models to output a calibrated probability of correctness. Our methods enable models to either abstain from producing a complete response or flag individual claims where uncertainty remains. Utilizing Qwen3-4B-Instruct, empirical analysis reveals behavior-calibrated reinforcement learning allows smaller models to surpass frontier models in uncertainty quantification--a transferable meta-skill decouplable from raw predictive accuracy. Trained on math reasoning tasks, our model's log-scale Accuracy-to-Hallucination Ratio gain (0.806) exceeds GPT-5's (0.207) in a challenging in-domain evaluation (BeyondAIME). Moreover, in cross-domain factual QA (SimpleQA), our 4B LLM achieves zero-shot calibration error on par with frontier models including Grok-4 and Gemini-2.5-Pro, even though its factual accuracy is much lower.
Authors: Rui Pan, Zhuofu Chen, Hongyi Liu, Arvind Krishnamurthy, Ravi Netravali
Abstract: Diffusion Large Language Models (dLLMs) offer fast, parallel token generation, but their standalone use is plagued by an inherent efficiency-quality tradeoff. We show that, if carefully applied, the attributes of dLLMs can actually be a strength for drafters in speculative decoding with autoregressive (AR) verifiers. Our core insight is that dLLM's speed from parallel decoding drastically lowers the risk of costly rejections, providing a practical mechanism to effectively realize the (elusive) lengthy drafts that lead to large speedups with speculative decoding. We present FailFast, a dLLM-based speculative decoding framework that realizes this approach by dynamically adapting its speculation length. It "fails fast" by spending minimal compute in hard-to-speculate regions to shrink speculation latency and "wins big" by aggressively extending draft lengths in easier regions to reduce verification latency (in many cases, speculating and accepting 70 tokens at a time!). Without any fine-tuning, FailFast delivers lossless acceleration of AR LLMs and achieves up to 4.9$\times$ speedup over vanilla decoding, 1.7$\times$ over the best naive dLLM drafter, and 1.7$\times$ over EAGLE-3 across diverse models and workloads. We open-source FailFast at https://github.com/ruipeterpan/failfast.
Authors: Naman Agarwal, Siddhartha R. Dalal, Vishal Misra
Abstract: Transformers often appear to perform Bayesian reasoning in context, but verifying this rigorously has been impossible: natural data lack analytic posteriors, and large models conflate reasoning with memorization. We address this by constructing \emph{Bayesian wind tunnels} -- controlled environments where the true posterior is known in closed form and memorization is provably impossible. In these settings, small transformers reproduce Bayesian posteriors with $10^{-3}$-$10^{-4}$ bit accuracy, while capacity-matched MLPs fail by orders of magnitude, establishing a clear architectural separation. Across two tasks -- bijection elimination and Hidden Markov Model (HMM) state tracking -- we find that transformers implement Bayesian inference through a consistent geometric mechanism: residual streams serve as the belief substrate, feed-forward networks perform the posterior update, and attention provides content-addressable routing. Geometric diagnostics reveal orthogonal key bases, progressive query-key alignment, and a low-dimensional value manifold parameterized by posterior entropy. During training this manifold unfurls while attention patterns remain stable, a \emph{frame-precision dissociation} predicted by recent gradient analyses. Taken together, these results demonstrate that hierarchical attention realizes Bayesian inference by geometric design, explaining both the necessity of attention and the failure of flat architectures. Bayesian wind tunnels provide a foundation for mechanistically connecting small, verifiable systems to reasoning phenomena observed in large language models.
Authors: Naman Agarwal, Siddhartha R. Dalal, Vishal Misra
Abstract: Transformers empirically perform precise probabilistic reasoning in carefully constructed ``Bayesian wind tunnels'' and in large-scale language models, yet the mechanisms by which gradient-based learning creates the required internal geometry remain opaque. We provide a complete first-order analysis of how cross-entropy training reshapes attention scores and value vectors in a transformer attention head. Our core result is an \emph{advantage-based routing law} for attention scores, \[ \frac{\partial L}{\partial s_{ij}} = \alpha_{ij}\bigl(b_{ij}-\mathbb{E}_{\alpha_i}[b]\bigr), \qquad b_{ij} := u_i^\top v_j, \] coupled with a \emph{responsibility-weighted update} for values, \[ \Delta v_j = -\eta\sum_i \alpha_{ij} u_i, \] where $u_i$ is the upstream gradient at position $i$ and $\alpha_{ij}$ are attention weights. These equations induce a positive feedback loop in which routing and content specialize together: queries route more strongly to values that are above-average for their error signal, and those values are pulled toward the queries that use them. We show that this coupled specialization behaves like a two-timescale EM procedure: attention weights implement an E-step (soft responsibilities), while values implement an M-step (responsibility-weighted prototype updates), with queries and keys adjusting the hypothesis frame. Through controlled simulations, including a sticky Markov-chain task where we compare a closed-form EM-style update to standard SGD, we demonstrate that the same gradient dynamics that minimize cross-entropy also sculpt the low-dimensional manifolds identified in our companion work as implementing Bayesian inference. This yields a unified picture in which optimization (gradient flow) gives rise to geometry (Bayesian manifolds), which in turn supports function (in-context probabilistic reasoning).
Authors: Dakuan Lu, Jiaqi Zhang, Cheng Yuan, Jiawei Shao, Xuelong Li
Abstract: Recent advances in large language models (LLMs) have been largely driven by scaling laws for individual models, which predict performance improvements as model parameters and data volume increase. However, the capabilities of any single LLM are inherently bounded. One solution originates from intricate interactions among multiple LLMs, rendering their collective performance surpasses that of any constituent model. Despite the rapid proliferation of multi-model integration techniques such as model routing and post-hoc ensembling, a unifying theoretical framework of performance scaling for multi-model collaboration remains absent. In this work, we propose the Law of Multi-model Collaboration, a scaling law that predicts the performance limits of LLM ensembles based on their aggregated parameter budget. To quantify the intrinsic upper bound of multi-model collaboration, we adopt a method-agnostic formulation and assume an idealized integration oracle where the total cross-entropy loss of each sample is determined by the minimum loss of any model in the model pool. Experimental results reveal that multi-model systems follow a power-law scaling with respect to the total parameter count, exhibiting a more significant improvement trend and a lower theoretical loss floor compared to single model scaling. Moreover, ensembles of heterogeneous model families achieve better performance scaling than those formed within a single model family, indicating that model diversity is a primary driver of collaboration gains. These findings suggest that model collaboration represents a critical axis for extending the intelligence frontier of LLMs.
Authors: Hanzheng Li, Xi Fang, Yixuan Li, Chaozheng Huang, Junjie Wang, Xi Wang, Hongzhe Bai, Bojun Hao, Shenyu Lin, Huiqi Liang, Linfeng Zhang, Guolin Ke
Abstract: The integration of Multimodal Large Language Models (MLLMs) into chemistry promises to revolutionize scientific discovery, yet their ability to comprehend the dense, graphical language of reactions within authentic literature remains underexplored. Here, we introduce RxnBench, a multi-tiered benchmark designed to rigorously evaluate MLLMs on chemical reaction understanding from scientific PDFs. RxnBench comprises two tasks: Single-Figure QA (SF-QA), which tests fine-grained visual perception and mechanistic reasoning using 1,525 questions derived from 305 curated reaction schemes, and Full-Document QA (FD-QA), which challenges models to synthesize information from 108 articles, requiring cross-modal integration of text, schemes, and tables. Our evaluation of MLLMs reveals a critical capability gap: while models excel at extracting explicit text, they struggle with deep chemical logic and precise structural recognition. Notably, models with inference-time reasoning significantly outperform standard architectures, yet none achieve 50\% accuracy on FD-QA. These findings underscore the urgent need for domain-specific visual encoders and stronger reasoning engines to advance autonomous AI chemists.
Authors: Naman Agarwal, Siddhartha R. Dalal, Vishal Misra
Abstract: Recent work has shown that small transformers trained in controlled "wind-tunnel'' settings can implement exact Bayesian inference, and that their training dynamics produce a geometric substrate -- low-dimensional value manifolds and progressively orthogonal keys -- that encodes posterior structure. We investigate whether this geometric signature persists in production-grade language models. Across Pythia, Phi-2, Llama-3, and Mistral families, we find that last-layer value representations organize along a single dominant axis whose position strongly correlates with predictive entropy, and that domain-restricted prompts collapse this structure into the same low-dimensional manifolds observed in synthetic settings. To probe the role of this geometry, we perform targeted interventions on the entropy-aligned axis of Pythia-410M during in-context learning. Removing or perturbing this axis selectively disrupts the local uncertainty geometry, whereas matched random-axis interventions leave it intact. However, these single-layer manipulations do not produce proportionally specific degradation in Bayesian-like behavior, indicating that the geometry is a privileged readout of uncertainty rather than a singular computational bottleneck. Taken together, our results show that modern language models preserve the geometric substrate that enables Bayesian inference in wind tunnels, and organize their approximate Bayesian updates along this substrate.
Authors: Chaodong Tong, Qi Zhang, Chen Li, Lei Jiang, Yanbing Liu
Abstract: Faithfulness hallucinations in VQA occur when vision-language models produce fluent yet visually ungrounded answers, severely undermining their reliability in safety-critical applications. Existing detection methods mainly fall into two categories: external verification approaches relying on auxiliary models or knowledge bases, and uncertainty-driven approaches using repeated sampling or uncertainty estimates. The former suffer from high computational overhead and are limited by external resource quality, while the latter capture only limited facets of model uncertainty and fail to sufficiently explore the rich internal signals associated with the diverse failure modes. Both paradigms thus have inherent limitations in efficiency, robustness, and detection performance. To address these challenges, we propose FaithSCAN: a lightweight network that detects hallucinations by exploiting rich internal signals of VLMs, including token-level decoding uncertainty, intermediate visual representations, and cross-modal alignment features. These signals are fused via branch-wise evidence encoding and uncertainty-aware attention. We also extend the LLM-as-a-Judge paradigm to VQA hallucination and propose a low-cost strategy to automatically generate model-dependent supervision signals, enabling supervised training without costly human labels while maintaining high detection accuracy. Experiments on multiple VQA benchmarks show that FaithSCAN significantly outperforms existing methods in both effectiveness and efficiency. In-depth analysis shows hallucinations arise from systematic internal state variations in visual perception, cross-modal reasoning, and language decoding. Different internal signals provide complementary diagnostic cues, and hallucination patterns vary across VLM architectures, offering new insights into the underlying causes of multimodal hallucinations.
Authors: Ching Ho Lee, Javier Nistal, Stefan Lattner, Marco Pasini, George Fazekas
Abstract: We study timbre transfer as an inference-time editing problem for music audio. Starting from a strong pre-trained latent diffusion model, we introduce a lightweight procedure that requires no additional training: (i) a dimension-wise noise injection that targets latent channels most informative of instrument identity, and (ii) an early-step clamping mechanism that re-imposes the input's melodic and rhythmic structure during reverse diffusion. The method operates directly on audio latents and is compatible with text/audio conditioning (e.g., CLAP). We discuss design choices,analyze trade-offs between timbral change and structural preservation, and show that simple inference-time controls can meaningfully steer pre-trained models for style-transfer use cases.
Authors: Guanyuan Pan, Shuai Wang, Yugui Lin, Tiansheng Zhou, Pietro Li\`o, Zhenxin Zhao, Yaqi Wang
Abstract: Analog mixed-signal circuit sizing involves complex trade-offs within high-dimensional design spaces. Existing automatic analog circuit sizing approaches rely solely on netlists, ignoring the circuit schematic, which hinders the cognitive link between the schematic and its performance. Furthermore, the black-box nature of machine learning methods and hallucination risks in large language models fail to provide the necessary ground-truth explainability required for industrial sign-off. To address these challenges, we propose a Vision Language Model-optimized collaborative agent design workflow (VLM-CAD), which analyzes circuits, optimizes DC operating points, performs inference-based sizing, and executes external sizing optimization. We integrate Image2Net to annotate circuit schematics and generate a structured JSON description for precise interpretation by Vision Language Models. Furthermore, we propose an Explainable Trust Region Bayesian Optimization method (ExTuRBO) that employs collaborative warm-start from agent-generated seeds and offers dual-granularity sensitivity analysis for external sizing optimization, supporting a comprehensive final design report. Experiment results on amplifier sizing tasks using 180nm, 90nm, and 45nm Predictive Technology Models demonstrate that VLM-CAD effectively balances power and performance while maintaining physics-based explainability. VLM-CAD meets all specification requirements while maintaining low power consumption in optimizing an amplifier with a complementary input and a class-AB output stage, with a total runtime under 66 minutes across all experiments on two amplifiers.
Authors: Yuan Cheng, Fengzhuo Zhang, Yunlong Hou, Cunxiao Du, Chao Du, Tianyu Pang, Aixin Sun, Zhuoran Yang
Abstract: Large Language Models (LLMs) often exhibit slash attention patterns, where attention scores concentrate along the $\Delta$-th sub-diagonal for some offset $\Delta$. These patterns play a key role in passing information across tokens. But why do they emerge? In this paper, we demystify the emergence of these Slash-Dominant Heads (SDHs) from both empirical and theoretical perspectives. First, by analyzing open-source LLMs, we find that SDHs are intrinsic to models and generalize to out-of-distribution prompts. To explain the intrinsic emergence, we analyze the queries, keys, and Rotary Position Embedding (RoPE), which jointly determine attention scores. Our empirical analysis reveals two characteristic conditions of SDHs: (1) Queries and keys are almost rank-one, and (2) RoPE is dominated by medium- and high-frequency components. Under these conditions, queries and keys are nearly identical across tokens, and interactions between medium- and high-frequency components of RoPE give rise to SDHs. Beyond empirical evidence, we theoretically show that these conditions are sufficient to ensure the emergence of SDHs by formalizing them as our modeling assumptions. Particularly, we analyze the training dynamics of a shallow Transformer equipped with RoPE under these conditions, and prove that models trained via gradient descent exhibit SDHs. The SDHs generalize to out-of-distribution prompts.
Authors: Ziang Cui, Mengran Yu, Tianjiao Li, Chenyu Shi, Yingxuan Shi, Lusheng Zhang, Hongwei Lin
Abstract: Large Language Models (LLMs) have achieved remarkable strides in multilingual translation but are hindered by a systemic cross-lingual verbosity bias, rendering them unsuitable for strict time-constrained tasks like subtitling and dubbing. Current prompt-engineering approaches struggle to resolve this conflict between semantic fidelity and rigid temporal feasibility. To bridge this gap, we first introduce Sand-Glass, a benchmark specifically designed to evaluate translation under syllable-level duration constraints. Furthermore, we propose HOMURA, a reinforcement learning framework that explicitly optimizes the trade-off between semantic preservation and temporal compliance. By employing a KL-regularized objective with a novel dynamic syllable-ratio reward, HOMURA effectively "tames" the output length. Experimental results demonstrate that our method significantly outperforms strong LLM baselines, achieving precise length control that respects linguistic density hierarchies without compromising semantic adequacy.
Authors: Jianshu She, Zonghang Li, Hongchao Du, Shangyu Wu, Wenhao Zheng, Eric Xing, Zhengzhong Liu, Huaxiu Yao, Jason Xue, Qirong Ho
Abstract: LAPS identifies and disaggregates requests with different prompt lengths in LLM serving to reduce TTFT latency. While recent systems have decoupled the prefill and decode stages to improve throughput, they still rely on unified scheduling policies that fail to adapt to heterogeneous workload characteristics. We observe that prompt-length variations lead to distinct performance bottlenecks, motivating an adaptive scheduling strategy. LAPS disaggregates multi-turn long-prefill requests from short-prefill ones and introduces a length-aware smart batching mechanism for short-prefill workloads. It adopts a dual-queue design that supports temporal disaggregation on a single prefill instance or spatial disaggregation across multiple instances. For short-prefill batches, a batch waiting window and CUDA Graph-based clustering mitigate interference from heterogeneous computation, reducing batching delay and lowering average latency. In real multi-turn workloads, LAPS reduces prefill latency by over 30\% compared to vanilla SGLang under prefill-decode disaggregation, and further decreases SLO violations by 28\% in multi-instance deployments with vanilla data-parallel configuration. Compared to the SGLang router with load balancing, it further lowers SLO violations by 12\% in multi-GPU settings. Under high concurrency and mixed-request scenarios, LAPS improves request throughput by 35\% serving Qwen2.5-32B model for prefill instance, demonstrating its effectiveness in optimizing heterogeneous LLM serving workloads.
Authors: V\'ictor Yeste, Paolo Rosso
Abstract: We study sentence-level detection of the 19 human values in the refined Schwartz continuum in about 74k English sentences from news and political manifestos (ValueEval'24 corpus). Each sentence is annotated with value presence, yielding a binary moral-presence label and a 19-way multi-label task under severe class imbalance. First, we show that moral presence is learnable from single sentences: a DeBERTa-base classifier attains positive-class F1 = 0.74 with calibrated thresholds. Second, we compare direct multi-label value detectors with presence-gated hierarchies under a single 8 GB GPU budget. Under matched compute, presence gating does not improve over direct prediction, indicating that gate recall becomes a bottleneck. Third, we investigate lightweight auxiliary signals - short-range context, LIWC-22 and moral lexica, and topic features - and small ensembles. Our best supervised configuration, a soft-voting ensemble of DeBERTa-based models enriched with such signals, reaches macro-F1 = 0.332 on the 19 values, improving over the best previous English-only baseline on this corpus (macro-F1 $\approx$ 0.28). We additionally benchmark 7-9B instruction-tuned LLMs (Gemma 2 9B, Llama 3.1 8B, Mistral 8B, Qwen 2.5 7B) in zero-/few-shot and QLoRA setups, and find that they lag behind the supervised ensemble under the same hardware constraint. Overall, our results provide empirical guidance for building compute-efficient, value-aware NLP models under realistic GPU budgets.
Authors: Yuzhen Shi, Huanghai Liu, Yiran Hu, Gaojie Song, Xinran Xu, Yubo Ma, Tianyi Tang, Li Zhang, Qingjing Chen, Di Feng, Wenbo Lv, Weiheng Wu, Kexin Yang, Sen Yang, Wei Wang, Rongyao Shi, Yuanyang Qiu, Yuemeng Qi, Jingwen Zhang, Xiaoyu Sui, Yifan Chen, Yi Zhang, An Yang, Bowen Yu, Dayiheng Liu, Junyang Lin, Weixing Shen, Bing Zhao, Charles L. A. Clarke, Hu Wei
Abstract: As large language models (LLMs) are increasingly applied to legal domain-specific tasks, evaluating their ability to perform legal work in real-world settings has become essential. However, existing legal benchmarks rely on simplified and highly standardized tasks, failing to capture the ambiguity, complexity, and reasoning demands of real legal practice. Moreover, prior evaluations often adopt coarse, single-dimensional metrics and do not explicitly assess fine-grained legal reasoning. To address these limitations, we introduce PLawBench, a Practical Law Benchmark designed to evaluate LLMs in realistic legal practice scenarios. Grounded in real-world legal workflows, PLawBench models the core processes of legal practitioners through three task categories: public legal consultation, practical case analysis, and legal document generation. These tasks assess a model's ability to identify legal issues and key facts, perform structured legal reasoning, and generate legally coherent documents. PLawBench comprises 850 questions across 13 practical legal scenarios, with each question accompanied by expert-designed evaluation rubrics, resulting in approximately 12,500 rubric items for fine-grained assessment. Using an LLM-based evaluator aligned with human expert judgments, we evaluate 10 state-of-the-art LLMs. Experimental results show that none achieves strong performance on PLawBench, revealing substantial limitations in the fine-grained legal reasoning capabilities of current LLMs and highlighting important directions for future evaluation and development of legal LLMs. Data is available at: https://github.com/skylenage/PLawbench.
Authors: Longteng Zhang, Sen Wu, Shuai Hou, Zhengyu Qing, Zhuo Zheng, Danning Ke, Qihong Lin, Qiang Wang, Shaohuai Shi, Xiaowen Chu
Abstract: Adapting large pre-trained language models to downstream tasks often entails fine-tuning millions of parameters or deploying costly dense weight updates, which hinders their use in resource-constrained environments. Low-rank Adaptation (LoRA) reduces trainable parameters by factorizing weight updates, yet the underlying dense weights still impose high storage and computation costs. Magnitude-based pruning can yield sparse models but typically degrades LoRA's performance when applied naively. In this paper, we introduce SALR (Sparsity-Aware Low-Rank Representation), a novel fine-tuning paradigm that unifies low-rank adaptation with sparse pruning under a rigorous mean-squared-error framework. We prove that statically pruning only the frozen base weights minimizes the pruning error bound, and we recover the discarded residual information via a truncated-SVD low-rank adapter, which provably reduces per-entry MSE by a factor of $(1 - r/\min(d,k))$. To maximize hardware efficiency, we fuse multiple low-rank adapters into a single concatenated GEMM, and we adopt a bitmap-based encoding with a two-stage pipelined decoding + GEMM design to achieve true model compression and speedup. Empirically, SALR attains 50\% sparsity on various LLMs while matching the performance of LoRA on GSM8K and MMLU, reduces model size by $2\times$, and delivers up to a $1.7\times$ inference speedup.
Authors: Preethi Seshadri, Samuel Cahyawijaya, Ayomide Odumakinde, Sameer Singh, Seraphina Goldfarb-Tarrant
Abstract: Agentic benchmarks increasingly rely on LLM-simulated users to scalably evaluate agent performance, yet the robustness, validity, and fairness of this approach remain unexamined. Through a user study with participants across the United States, India, Kenya, and Nigeria, we investigate whether LLM-simulated users serve as reliable proxies for real human users in evaluating agents on {\tau}-Bench retail tasks. We find that user simulation lacks robustness, with agent success rates varying up to 9 percentage points across different user LLMs. Furthermore, evaluations using simulated users exhibit systematic miscalibration, underestimating agent performance on challenging tasks and overestimating it on moderately difficult ones. African American Vernacular English (AAVE) speakers experience consistently worse success rates and calibration errors than Standard American English (SAE) speakers, with disparities compounding significantly with age. We also find simulated users to be a differentially effective proxy for different populations, performing worst for AAVE and Indian English speakers. Additionally, simulated users introduce conversational artifacts and surface different failure patterns than human users. These findings demonstrate that current evaluation practices risk misrepresenting agent capabilities across diverse user populations and may obscure real-world deployment challenges.
Authors: Murat Arda Onsu, Poonam Lohan, Burak Kantarci, Aisha Syed, Matthew Andrews, Sean Kennedy
Abstract: Intelligent Transportation Systems (ITS) demand real-time collision prediction to ensure road safety and reduce accident severity. Conventional approaches rely on transmitting raw video or high-dimensional sensory data from roadside units (RSUs) to vehicles, which is impractical under vehicular communication bandwidth and latency constraints. In this work, we propose a semantic V2X framework in which RSU-mounted cameras generate spatiotemporal semantic embeddings of future frames using the Video Joint Embedding Predictive Architecture (V-JEPA). To evaluate the system, we construct a digital twin of an urban traffic environment enabling the generation of d verse traffic scenarios with both safe and collision events. These embeddings of the future frame, extracted from V-JEPA, capture task-relevant traffic dynamics and are transmitted via V2X links to vehicles, where a lightweight attentive probe and classifier decode them to predict imminent collisions. By transmitting only semantic embeddings instead of raw frames, the proposed system significantly reduces communication overhead while maintaining predictive accuracy. Experimental results demonstrate that the framework with an appropriate processing method achieves a 10% F1-score improvement for collision prediction while reducing transmission requirements by four orders of magnitude compared to raw video. This validates the potential of semantic V2X communication to enable cooperative, real-time collision prediction in ITS.
Authors: Zecheng Tang, Quantong Qiu, Yi Yang, Zhiyi Hong, Haiya Xiang, Kebin Liu, Qingqing Dang, Juntao Li, Min Zhang
Abstract: The quadratic complexity of standard attention mechanisms poses a significant scalability bottleneck for large language models (LLMs) in long-context scenarios. While hybrid attention strategies that combine sparse and full attention within a single model offer a viable solution, they typically employ static computation ratios (i.e., fixed proportions of sparse versus full attention) and fail to adapt to the varying sparsity sensitivities of downstream tasks during inference. To address this issue, we propose Elastic Attention, which allows the model to dynamically adjust its overall sparsity based on the input. This is achieved by integrating a lightweight Attention Router into the existing pretrained model, which dynamically assigns each attention head to different computation modes. Within only 12 hours of training on 8xA800 GPUs, our method enables models to achieve both strong performance and efficient inference. Experiments across three long-context benchmarks on widely-used LLMs demonstrate the superiority of our method.
Authors: Vi Vu, Thanh-Huy Nguyen, Tien-Thinh Nguyen, Ba-Thinh Lam, Hoang-Thien Nguyen, Tianyang Wang, Xingjian Li, Min Xu
Abstract: Foundation models like the Segment Anything Model (SAM) show strong generalization, yet adapting them to medical images remains difficult due to domain shift, scarce labels, and the inability of Parameter-Efficient Fine-Tuning (PEFT) to exploit unlabeled data. While conventional models like U-Net excel in semi-supervised medical learning, their potential to assist a PEFT SAM has been largely overlooked. We introduce SC-SAM, a specialist-generalist framework where U-Net provides point-based prompts and pseudo-labels to guide SAM's adaptation, while SAM serves as a powerful generalist supervisor to regularize U-Net. This reciprocal guidance forms a bidirectional co-training loop that allows both models to effectively exploit the unlabeled data. Across prostate MRI and polyp segmentation benchmarks, our method achieves state-of-the-art results, outperforming other existing semi-supervised SAM variants and even medical foundation models like MedSAM, highlighting the value of specialist-generalist cooperation for label-efficient medical image segmentation. Our code is available at https://github.com/vnlvi2k3/SC-SAM.
Authors: Ren Zhuang, Ben Wang, Shuifa Sun
Abstract: Scaling test-time compute enhances long chain-of-thought (CoT) reasoning, yet existing approaches face a fundamental trade-off between computational cost and coverage quality: either incurring high training expense or yielding redundant trajectories. We introduce The Geometric Reasoner (TGR), a training-free framework that performs manifold-informed latent foresight search under strict memory bounds. At each chunk boundary, TGR scores candidate latent anchors via a lightweight look-ahead estimate combined with soft geometric regularizers that encourage smooth trajectories and diverse exploration. Chunk-wise KV cache resets keep memory linear in chunk length. On challenging math and code benchmarks, TGR improves robust trajectory coverage, measured by the area under the Pass@$k$ curve (AUC), by up to 13 points on Qwen3-8B, with negligible overhead of about 1.1--1.3 times.
Authors: Oren Sultan, Jordi Armengol-Estape, Pascal Kesseli, Julien Vanegue, Dafna Shahaf, Yossi Adi, Peter O'Hearn
Abstract: Determining whether a program terminates is a central problem in computer science. Turing's foundational result established the Halting Problem as undecidable, showing that no algorithm can universally determine termination for all programs and inputs. Consequently, automatic verification tools approximate termination, sometimes failing to prove or disprove; these tools rely on problem-specific architectures and abstractions, and are usually tied to particular programming languages. Recent success and progress in large language models (LLMs) raises the following question: can LLMs reliably predict program termination? In this work, we evaluate LLMs on a diverse set of C programs from the Termination category of the International Competition on Software Verification (SV-Comp) 2025. Our results suggest that LLMs perform remarkably well at predicting program termination, where GPT-5 and Claude Sonnet-4.5 would rank just behind the top-ranked tool (using test-time-scaling), and Code World Model (CWM) would place just behind the second-ranked tool. While LLMs are effective at predicting program termination, they often fail to provide a valid witness as a proof. Moreover, LLMs performance drops as program length increases. We hope these insights motivate further research into program termination and the broader potential of LLMs for reasoning about undecidable problems.
Authors: Antanas Zilinskas, Robert N. Shorten, Jakub Marecek
Abstract: Forecasting rare events in multivariate time-series data is challenging due to severe class imbalance, long-range dependencies, and distributional uncertainty. We introduce EVEREST, a transformer-based architecture for probabilistic rare-event forecasting that delivers calibrated predictions and tail-aware risk estimation, with auxiliary interpretability via attention-based signal attribution. EVEREST integrates four components: (i) a learnable attention bottleneck for soft aggregation of temporal dynamics; (ii) an evidential head for estimating aleatoric and epistemic uncertainty via a Normal--Inverse--Gamma distribution; (iii) an extreme-value head that models tail risk using a Generalized Pareto Distribution; and (iv) a lightweight precursor head for early-event detection. These modules are jointly optimized with a composite loss (focal loss, evidential NLL, and a tail-sensitive EVT penalty) and act only at training time; deployment uses a single classification head with no inference overhead (approximately 0.81M parameters). On a decade of space-weather data, EVEREST achieves state-of-the-art True Skill Statistic (TSS) of 0.973/0.970/0.966 at 24/48/72-hour horizons for C-class flares. The model is compact, efficient to train on commodity hardware, and applicable to high-stakes domains such as industrial monitoring, weather, and satellite diagnostics. Limitations include reliance on fixed-length inputs and exclusion of image-based modalities, motivating future extensions to streaming and multimodal forecasting.
Authors: Puzhen Wu, Han Weng, Quan Zheng, Yi Zhan, Hewei Wang, Yiming Li, Jiahui Han, Rui Xu
Abstract: Improper exposure often leads to severe loss of details, color distortion, and reduced contrast. Exposure correction still faces two critical challenges: (1) the ignorance of object-wise regional semantic information causes the color shift artifacts; (2) real-world exposure images generally have no ground-truth labels, and its labeling entails massive manual editing. To tackle the challenges, we propose a new unsupervised semantic-aware exposure correction network. It contains an adaptive semantic-aware fusion module, which effectively fuses the semantic information extracted from a pre-trained Fast Segment Anything Model into a shared image feature space. Then the fused features are used by our multi-scale residual spatial mamba group to restore the details and adjust the exposure. To avoid manual editing, we propose a pseudo-ground truth generator guided by CLIP, which is fine-tuned to automatically identify exposure situations and instruct the tailored corrections. Also, we leverage the rich priors from the FastSAM and CLIP to develop a semantic-prompt consistency loss to enforce semantic consistency and image-prompt alignment for unsupervised training. Comprehensive experimental results illustrate the effectiveness of our method in correcting real-world exposure images and outperforms state-of-the-art unsupervised methods both numerically and visually.
Authors: Jinkyu Sung, Myunggeum Jee, Joonseok Lee
Abstract: Link sign prediction on a signed graph is a task to determine whether the relationship represented by an edge is positive or negative. Since the presence of negative edges violates the graph homophily assumption that adjacent nodes are similar, regular graph methods have not been applicable without auxiliary structures to handle them. We aim to directly model the latent statistical dependency among edges with the Gaussian copula and its corresponding correlation matrix, extending CopulaGNN (Ma et al., 2021). However, a naive modeling of edge-edge relations is computationally intractable even for a graph with moderate scale. To address this, we propose to 1) represent the correlation matrix as a Gramian of edge embeddings, significantly reducing the number of parameters, and 2) reformulate the conditional probability distribution to dramatically reduce the inference cost. We theoretically verify scalability of our method by proving its linear convergence. Also, our extensive experiments demonstrate that it achieves significantly faster convergence than baselines, maintaining competitive prediction performance to the state-of-the-art models.
Authors: Kaehyun Um, KyuHwan Yeom, Haerim Yang, Minyoung Choi, Hyeongjun Yang, Kyong-Ho Lee
Abstract: Large Language Models (LLMs) have recently demonstrated remarkable reasoning abilities, yet hallucinate on knowledge-intensive tasks. Retrieval-augmented generation (RAG) mitigates this issue by grounding answers in external sources, e.g., knowledge graphs (KGs). However, existing KG-based RAG approaches rely on semantics-unaware path sampling and are weakly aligned with KG reasoning objectives, which limits further accuracy gains. They also feed retrieved paths directly into the reasoner without organizing them into answer-centered reasoning paths, hindering small LLMs' ability to leverage the retrieved knowledge. Furthermore, prior works predominantly rely on large LLMs (e.g., ChatGPT/GPT-4) or assume backbones above 7B parameters, leaving sub-7B models underexplored. We address this gap with RPO-RAG, the first KG-based RAG framework specifically designed for small LLMs, to the best of our knowledge. RPO-RAG introduces three key innovations: (1) a query-path semantic sampling strategy that provides informative supervisory signals; (2) a relation-aware preference optimization that aligns training with intermediate KG reasoning signals (e.g., relation); and (3) an answer-centered prompt design that organizes entities and reasoning paths in an interpretable format. Extensive experiments on two benchmark Knowledge Graph Question Answering (KGQA) datasets, WebQSP and CWQ, demonstrate that RPO-RAG effectively bridges the performance gap between small and large language models. On WebQSP, it improves F1 by up to 8.8%, reflecting enhanced answer precision, while on CWQ it achieves new state-of-the-art results among models under 8B parameters in both Hit and F1. Overall, RPO-RAG substantially improves the reasoning capability of small LLMs, even under 3B parameters-highlighting their potential for resource-efficient and practical on-device KGQA applications.
Authors: Fakrul Islam Tushar, Joseph Y. Lo
Abstract: Using multiple open-access models trained on public datasets, we developed Tri-Reader, a comprehensive, freely available pipeline that integrates lung segmentation, nodule detection, and malignancy classification into a unified tri-stage workflow. The pipeline is designed to prioritize sensitivity while reducing the candidate burden for annotators. To ensure accuracy and generalizability across diverse practices, we evaluated Tri-Reader on multiple internal and external datasets as compared with expert annotations and dataset-provided reference standards.
Authors: Zhizheng Jiang, Kang Zhao, Weikai Xu, Xinkui Lin, Wei Liu, Jian Luan, Shuo Shang, Peng Han
Abstract: Large reasoning models (LRMs) aim to solve diverse and complex problems through structured reasoning. Recent advances in group-based policy optimization methods have shown promise in enabling stable advantage estimation without reliance on process-level annotations. However, these methods rely on advantage gaps induced by high-quality samples within the same batch, which makes the training process fragile and inefficient when intra-group advantages collapse under challenging tasks. To address these problems, we propose a reinforcement learning mechanism named \emph{\textbf{R^3}} that along three directions: (1) a \emph{cross-context \underline{\textbf{R}}eplay} strategy that maintains the intra-group advantage by recalling valuable examples from historical trajectories of the same query, (2) an \emph{in-context self-\underline{\textbf{R}}eflection} mechanism enabling models to refine outputs by leveraging past failures, and (3) a \emph{structural entropy \underline{\textbf{R}}anking reward}, which assigns relative rewards to truncated or failed samples by ranking responses based on token-level entropy patterns, capturing both local exploration and global stability. We implement our method on Deepseek-R1-Distill-Qwen-1.5B and train it on the DeepscaleR-40k in the math domain. Experiments demonstrate our method achieves SoTA performance on several math benchmarks, representing significant improvements and fewer reasoning tokens over the base models. Code and model will be released.
Authors: Waris Gill, Ahmad Humayun, Ali Anwar, Muhammad Ali Gulzar
Abstract: Federated Learning (FL) enables collaborative training of Large Language Models (LLMs) across distributed data sources while preserving privacy. However, when federated LLMs are deployed in critical applications, it remains unclear which client(s) contributed to specific generated responses, hindering debugging, malicious client identification, fair reward allocation, and trust verification. We present ProToken, a novel Provenance methodology for Token-level attribution in federated LLMs that addresses client attribution during autoregressive text generation while maintaining FL privacy constraints. ProToken leverages two key insights to enable provenance at each token: (1) transformer architectures concentrate task-specific signals in later blocks, enabling strategic layer selection for computational tractability, and (2) gradient-based relevance weighting filters out irrelevant neural activations, focusing attribution on neurons that directly influence token generation. We evaluate ProToken across 16 configurations spanning four LLM architectures (Gemma, Llama, Qwen, SmolLM) and four domains (medical, financial, mathematical, coding). ProToken achieves 98% average attribution accuracy in correctly localizing responsible client(s), and maintains high accuracy when the number of clients are scaled, validating its practical viability for real-world deployment settings.
Authors: Peter Zeng, Weiling Li, Amie Paige, Zhengxiang Wang, Panagiotis Kaliosis, Dimitris Samaras, Gregory Zelinsky, Susan Brennan, Owen Rambow
Abstract: For generative AI agents to partner effectively with human users, the ability to accurately predict human intent is critical. But this ability to collaborate remains limited by a critical deficit: an inability to model common ground. Here, we present a referential communication experiment with a factorial design involving director-matcher pairs (human-human, human-AI, AI-human, and AI-AI) that interact with multiple turns in repeated rounds to match pictures of objects not associated with any obvious lexicalized labels. We release the online pipeline for data collection, the tools and analyses for accuracy, efficiency, and lexical overlap, and a corpus of 356 dialogues (89 pairs over 4 rounds each) that unmasks LVLMs' limitations in interactively resolving referring expressions, a crucial skill that underlies human language use.