Authors: Kangyu Wang
Abstract: Machine Learning ML agents have been increasingly used in decision-making across a wide range of tasks and environments. These ML agents are typically designed to balance multiple objectives when making choices. Understanding how their decision-making processes align with or diverge from human reasoning is essential. Human agents often encounter hard choices, that is, situations where options are incommensurable; neither option is preferred, yet the agent is not indifferent between them. In such cases, human agents can identify hard choices and resolve them through deliberation. In contrast, current ML agents, due to fundamental limitations in Multi-Objective Optimisation or MOO methods, cannot identify hard choices, let alone resolve them. Neither Scalarised Optimisation nor Pareto Optimisation, the two principal MOO approaches, can capture incommensurability. This limitation generates three distinct alignment problems: the alienness of ML decision-making behaviour from a human perspective; the unreliability of preference-based alignment strategies for hard choices; and the blockage of alignment strategies pursuing multiple objectives. Evaluating two potential technical solutions, I recommend an ensemble solution that appears most promising for enabling ML agents to identify hard choices and mitigate alignment problems. However, no known technique allows ML agents to resolve hard choices through deliberation, as they cannot autonomously change their goals. This underscores the distinctiveness of human agency and urges ML researchers to reconceptualise machine autonomy and develop frameworks and methods that can better address this fundamental gap.
Authors: Yajie Yu, Yue Feng
Abstract: Multi-agents has exhibited significant intelligence in real-word simulations with Large language models (LLMs) due to the capabilities of social cognition and knowledge retrieval. However, existing research on agents equipped with effective cognition chains including reasoning, planning, decision-making and reflecting remains limited, especially in the dynamically interactive scenarios. In addition, unlike human, prompt-based responses face challenges in psychological state perception and empirical calibration during uncertain gaming process, which can inevitably lead to cognition bias. In light of above, we introduce PolicyEvol-Agent, a comprehensive LLM-empowered framework characterized by systematically acquiring intentions of others and adaptively optimizing irrational strategies for continual enhancement. Specifically, PolicyEvol-Agent first obtains reflective expertise patterns and then integrates a range of cognitive operations with Theory of Mind alongside internal and external perspectives. Simulation results, outperforming RL-based models and agent-based methods, demonstrate the superiority of PolicyEvol-Agent for final gaming victory. Moreover, the policy evolution mechanism reveals the effectiveness of dynamic guideline adjustments in both automatic and human evaluation.
Authors: Javier Fumanal-Idocin, Javier Andreu-Perez
Abstract: Classical machine learning classifiers tend to be overconfident can be unreliable outside of the laboratory benchmarks. Properly assessing the reliability of the output of the model per sample is instrumental for real-life scenarios where these systems are deployed. Because of this, different techniques have been employed to properly quantify the quality of prediction for a given model. These are most commonly Bayesian statistics and, more recently, conformal learning. Given a calibration set, conformal learning can produce outputs that are guaranteed to cover the target class with a desired significance level, and are more reliable than the standard confidence intervals used by Bayesian methods. In this work, we propose to use conformal learning with fuzzy rule-based systems in classification and show some metrics of their performance. Then, we discuss how the use of type 2 fuzzy sets can improve the quality of the output of the system compared to both fuzzy and crisp rules. Finally, we also discuss how the fine-tuning of the system can be adapted to improve the quality of the conformal prediction.
Authors: Junyoung Park, Dalton Jones, Matt Morse, Raghavv Goel, Mingu Lee, Chris Lott
Abstract: In this work, we demonstrate that distinctive keys during LLM inference tend to have high attention scores. We explore this phenomenon and propose KeyDiff, a training-free KV cache eviction method based on key similarity. This method facilitates the deployment of LLM-based application requiring long input prompts in resource-constrained environments with limited memory and compute budgets. Unlike other KV cache eviction methods, KeyDiff can process arbitrarily long prompts within strict resource constraints and efficiently generate responses. We demonstrate that KeyDiff computes the optimal solution to a KV cache selection problem that maximizes key diversity, providing a theoretical understanding of KeyDiff. Notably,KeyDiff does not rely on attention scores, allowing the use of optimized attention mechanisms like FlashAttention. We demonstrate the effectiveness of KeyDiff across diverse tasks and models, illustrating a performance gap of less than 0.04\% with 8K cache budget ($\sim$ 23\% KV cache reduction) from the non-evicting baseline on the LongBench benchmark for Llama 3.1-8B and Llama 3.2-3B.
Authors: Sarath Shekkizhar, Romain Cosentino
Abstract: This paper investigates multimodal agents, in particular, OpenAI's Computer-User Agent (CUA), trained to control and complete tasks through a standard computer interface, similar to humans. We evaluated the agent's performance on the New York Times Wordle game to elicit model behaviors and identify shortcomings. Our findings revealed a significant discrepancy in the model's ability to recognize colors correctly depending on the context. The model had a $5.36\%$ success rate over several hundred runs across a week of Wordle. Despite the immense enthusiasm surrounding AI agents and their potential to usher in Artificial General Intelligence (AGI), our findings reinforce the fact that even simple tasks present substantial challenges for today's frontier AI models. We conclude with a discussion of the potential underlying causes, implications for future development, and research directions to improve these AI systems.
Authors: Paresh Chaudhary, Yancheng Liang, Daphne Chen, Simon S. Du, Natasha Jaques
Abstract: Being able to cooperate with new people is an important component of many economically valuable AI tasks, from household robotics to autonomous driving. However, generalizing to novel humans requires training on data that captures the diversity of human behaviors. Adversarial training is one avenue for searching for such data and ensuring that agents are robust. However, it is difficult to apply in the cooperative setting because adversarial policies intentionally learn to sabotage the task instead of simulating valid cooperation partners. To address this challenge, we propose a novel strategy for overcoming self-sabotage that combines a pre-trained generative model to simulate valid cooperative agent policies with adversarial training to maximize regret. We call our method GOAT: Generative Online Adversarial Training. In this framework, the GOAT dynamically searches for and generates coordination strategies where the learning policy -- the Cooperator agent -- underperforms. GOAT enables better generalization by exposing the Cooperator to various challenging interaction scenarios. We maintain realistic coordination strategies by updating only the generative model's embedding while keeping its parameters frozen, thus avoiding adversarial exploitation. We evaluate GOAT with real human partners, and the results demonstrate state-of-the-art performance on the Overcooked benchmark, highlighting its effectiveness in generalizing to diverse human behaviors.
Authors: Jiayi Pan, Xiuyu Li, Long Lian, Charlie Snell, Yifei Zhou, Adam Yala, Trevor Darrell, Kurt Keutzer, Alane Suhr
Abstract: Scaling inference-time computation has substantially improved the reasoning capabilities of language models. However, existing methods have significant limitations: serialized chain-of-thought approaches generate overly long outputs, leading to increased latency and exhausted context windows, while parallel methods such as self-consistency suffer from insufficient coordination, resulting in redundant computations and limited performance gains. To address these shortcomings, we propose Adaptive Parallel Reasoning (APR), a novel reasoning framework that enables language models to orchestrate both serialized and parallel computations end-to-end. APR generalizes existing reasoning methods by enabling adaptive multi-threaded inference using spawn() and join() operations. A key innovation is our end-to-end reinforcement learning strategy, optimizing both parent and child inference threads to enhance task success rate without requiring predefined reasoning structures. Experiments on the Countdown reasoning task demonstrate significant benefits of APR: (1) higher performance within the same context window (83.4% vs. 60.0% at 4k context); (2) superior scalability with increased computation (80.1% vs. 66.6% at 20k total tokens); (3) improved accuracy at equivalent latency (75.2% vs. 57.3% at approximately 5,000ms). APR represents a step towards enabling language models to autonomously optimize their reasoning processes through adaptive allocation of computation.
Authors: Gengxian Cao, Fengyuan Li, Hong Duan, Ye Yang, Bofeng Wang, Donghe Li
Abstract: This paper introduces a novel multi-Agent framework that automates the end to end production of Qinqiang opera by integrating Large Language Models , visual generation, and Text to Speech synthesis. Three specialized agents collaborate in sequence: Agent1 uses an LLM to craft coherent, culturally grounded scripts;Agent2 employs visual generation models to render contextually accurate stage scenes; and Agent3 leverages TTS to produce synchronized, emotionally expressive vocal performances. In a case study on Dou E Yuan, the system achieved expert ratings of 3.8 for script fidelity, 3.5 for visual coherence, and 3.8 for speech accuracy-culminating in an overall score of 3.6, a 0.3 point improvement over a Single Agent baseline. Ablation experiments demonstrate that removing Agent2 or Agent3 leads to drops of 0.4 and 0.5 points, respectively, underscoring the value of modular collaboration. This work showcases how AI driven pipelines can streamline and scale the preservation of traditional performing arts, and points toward future enhancements in cross modal alignment, richer emotional nuance, and support for additional opera genres.
Authors: Md Millat, Md Motiur
Abstract: The current study describes a cost-effective method for adapting large language models (LLMs) for academic advising with study-abroad contexts in mind and for application in low-resource methods for acculturation. With the Mistral-7B-Instruct model applied with a Low-Rank Adaptation (LoRA) method and a 4-bit quantization method, the model underwent training in two distinct stages related to this study's purpose to enhance domain specificity while maintaining computational efficiency. In Phase 1, the model was conditioned with a synthetic dataset via the Gemini Pro API, and in Phase 2, it was trained with manually curated datasets from the StudyAbroadGPT project to achieve enhanced, contextualized responses. Technical innovations entailed memory-efficient quantization, parameter-efficient adaptation, and continuous training analytics via Weights & Biases. After training, this study demonstrated a reduction in training loss by 52.7%, 92% accuracy in domain-specific recommendations, achieved 95% markdown-based formatting support, and a median run-rate of 100 samples per second on off-the-shelf GPU equipment. These findings support the effective application of instruction-tuned LLMs within educational advisers, especially in low-resource institutional scenarios. Limitations included decreased generalizability and the application of a synthetically generated dataset, but this framework is scalable for adding new multilingual-augmented and real-time academic advising processes. Future directions may include plans for the integration of retrieval-augmented generation, applying dynamic quantization routines, and connecting to real-time academic databases to increase adaptability and accuracy.
Authors: Mir Md Sajid Sarwar, Rajarshi Ray
Abstract: Explaining unsolvability of planning problems is of significant research interest in Explainable AI Planning. AI planning literature has reported several research efforts on generating explanations of solutions to planning problems. However, explaining the unsolvability of planning problems remains a largely open and understudied problem. A widely practiced approach to plan generation and automated problem solving, in general, is to decompose tasks into sub-problems that help progressively converge towards the goal. In this paper, we propose to adopt the same philosophy of sub-problem identification as a mechanism for analyzing and explaining unsolvability of planning problems in hybrid systems. In particular, for a given unsolvable planning problem, we propose to identify common waypoints, which are universal obstacles to plan existence; in other words, they appear on every plan from the source to the planning goal. This work envisions such waypoints as sub-problems of the planning problem and the unreachability of any of these waypoints as an explanation for the unsolvability of the original planning problem. We propose a novel method of waypoint identification by casting the problem as an instance of the longest common subsequence problem, a widely popular problem in computer science, typically considered as an illustrative example for the dynamic programming paradigm. Once the waypoints are identified, we perform symbolic reachability analysis on them to identify the earliest unreachable waypoint and report it as the explanation of unsolvability. We present experimental results on unsolvable planning problems in hybrid domains.
Authors: Ning Wang, Zihan Yan, Weiyang Li, Chuan Ma, He Chen, Tao Xiang
Abstract: Embodied agents exhibit immense potential across a multitude of domains, making the assurance of their behavioral safety a fundamental prerequisite for their widespread deployment. However, existing research predominantly concentrates on the security of general large language models, lacking specialized methodologies for establishing safety benchmarks and input moderation tailored to embodied agents. To bridge this gap, this paper introduces a novel input moderation framework, meticulously designed to safeguard embodied agents. This framework encompasses the entire pipeline, including taxonomy definition, dataset curation, moderator architecture, model training, and rigorous evaluation. Notably, we introduce EAsafetyBench, a meticulously crafted safety benchmark engineered to facilitate both the training and stringent assessment of moderators specifically designed for embodied agents. Furthermore, we propose Pinpoint, an innovative prompt-decoupled input moderation scheme that harnesses a masked attention mechanism to effectively isolate and mitigate the influence of functional prompts on moderation tasks. Extensive experiments conducted on diverse benchmark datasets and models validate the feasibility and efficacy of the proposed approach. The results demonstrate that our methodologies achieve an impressive average detection accuracy of 94.58%, surpassing the performance of existing state-of-the-art techniques, alongside an exceptional moderation processing time of merely 0.002 seconds per instance.
Authors: Jie Zhu, Qian Chen, Huaixia Dou, Junhui Li, Lifan Guo, Feng Chen, Chi Zhang
Abstract: Effective reasoning remains a core challenge for large language models (LLMs) in the financial domain, where tasks often require domain-specific knowledge, precise numerical calculations, and strict adherence to compliance rules. We propose DianJin-R1, a reasoning-enhanced framework designed to address these challenges through reasoning-augmented supervision and reinforcement learning. Central to our approach is DianJin-R1-Data, a high-quality dataset constructed from CFLUE, FinQA, and a proprietary compliance corpus (Chinese Compliance Check, CCC), combining diverse financial reasoning scenarios with verified annotations. Our models, DianJin-R1-7B and DianJin-R1-32B, are fine-tuned from Qwen2.5-7B-Instruct and Qwen2.5-32B-Instruct using a structured format that generates both reasoning steps and final answers. To further refine reasoning quality, we apply Group Relative Policy Optimization (GRPO), a reinforcement learning method that incorporates dual reward signals: one encouraging structured outputs and another rewarding answer correctness. We evaluate our models on five benchmarks: three financial datasets (CFLUE, FinQA, and CCC) and two general reasoning benchmarks (MATH-500 and GPQA-Diamond). Experimental results show that DianJin-R1 models consistently outperform their non-reasoning counterparts, especially on complex financial tasks. Moreover, on the real-world CCC dataset, our single-call reasoning models match or even surpass the performance of multi-agent systems that require significantly more computational cost. These findings demonstrate the effectiveness of DianJin-R1 in enhancing financial reasoning through structured supervision and reward-aligned learning, offering a scalable and practical solution for real-world applications.
Authors: Anna Karnysheva, Christian Drescher, Dietrich Klakow
Abstract: As large language models (LLMs) become integral to intelligent user interfaces (IUIs), their role as decision-making agents raises critical concerns about alignment. Although extensive research has addressed issues such as factuality, bias, and toxicity, comparatively little attention has been paid to measuring alignment to preferences, i.e., the relative desirability of different alternatives, a concept used in decision making, economics, and social choice theory. However, a reliable decision-making agent makes choices that align well with user preferences. In this paper, we generalize existing methods that exploit LLMs for ranking alternative outcomes by addressing alignment with the broader and more flexible concept of user preferences, which includes both strict preferences and indifference among alternatives. To this end, we put forward design principles for using LLMs to implement rational choice functions, and provide the necessary tools to measure preference satisfaction. We demonstrate the applicability of our approach through an empirical study in a practical application of an IUI in the automotive domain.
Authors: Daocheng Fu, Zijun Chen, Renqiu Xia, Qi Liu, Yuan Feng, Hongbin Zhou, Renrui Zhang, Shiyang Feng, Peng Gao, Junchi Yan, Botian Shi, Bo Zhang, Yu Qiao
Abstract: Mathematical geometric problem solving (GPS) often requires effective integration of multimodal information and verifiable logical coherence. Despite the fast development of large language models in general problem solving, it remains unresolved regarding with both methodology and benchmarks, especially given the fact that exiting synthetic GPS benchmarks are often not self-verified and contain noise and self-contradicted information due to the illusion of LLMs. In this paper, we propose a scalable data engine called TrustGeoGen for problem generation, with formal verification to provide a principled benchmark, which we believe lays the foundation for the further development of methods for GPS. The engine synthesizes geometric data through four key innovations: 1) multimodal-aligned generation of diagrams, textual descriptions, and stepwise solutions; 2) formal verification ensuring rule-compliant reasoning paths; 3) a bootstrapping mechanism enabling complexity escalation via recursive state generation and 4) our devised GeoExplore series algorithms simultaneously produce multi-solution variants and self-reflective backtracking traces. By formal logical verification, TrustGeoGen produces GeoTrust-200K dataset with guaranteed modality integrity, along with GeoTrust-test testset. Experiments reveal the state-of-the-art models achieve only 49.17\% accuracy on GeoTrust-test, demonstrating its evaluation stringency. Crucially, models trained on GeoTrust achieve OOD generalization on GeoQA, significantly reducing logical inconsistencies relative to pseudo-label annotated by OpenAI-o1. Our code is available at https://github.com/Alpha-Innovator/TrustGeoGen
Authors: Siyu Zhou, Tianyi Zhou, Yijun Yang, Guodong Long, Deheng Ye, Jing Jiang, Chengqi Zhang
Abstract: Can we build accurate world models out of large language models (LLMs)? How can world models benefit LLM agents? The gap between the prior knowledge of LLMs and the specified environment's dynamics usually bottlenecks LLMs' performance as world models. To bridge the gap, we propose a training-free "world alignment" that learns an environment's symbolic knowledge complementary to LLMs. The symbolic knowledge covers action rules, knowledge graphs, and scene graphs, which are extracted by LLMs from exploration trajectories and encoded into executable codes to regulate LLM agents' policies. We further propose an RL-free, model-based agent "WALL-E 2.0" through the model-predictive control (MPC) framework. Unlike classical MPC requiring costly optimization on the fly, we adopt an LLM agent as an efficient look-ahead optimizer of future steps' actions by interacting with the neurosymbolic world model. While the LLM agent's strong heuristics make it an efficient planner in MPC, the quality of its planned actions is also secured by the accurate predictions of the aligned world model. They together considerably improve learning efficiency in a new environment. On open-world challenges in Mars (Minecraft like) and ALFWorld (embodied indoor environments), WALL-E 2.0 significantly outperforms existing methods, e.g., surpassing baselines in Mars by 16.1%-51.6% of success rate and by at least 61.7% in score. In ALFWorld, it achieves a new record 98% success rate after only 4 iterations.
Authors: Raquel Fernandez-Peralta, Javier Fumanal-Idocin, Javier Andreu-Perez
Abstract: Rule-based systems are a very popular form of explainable AI, particularly in the fuzzy community, where fuzzy rules are widely used for control and classification problems. However, fuzzy rule-based classifiers struggle to reach bigger traction outside of fuzzy venues, because users sometimes do not know about fuzzy and because fuzzy partitions are not so easy to interpret in some situations. In this work, we propose a methodology to reduce fuzzy rule-based classifiers to crisp rule-based classifiers. We study different possible crisp descriptions and implement an algorithm to obtain them. Also, we analyze the complexity of the resulting crisp classifiers. We believe that our results can help both fuzzy and non-fuzzy practitioners understand better the way in which fuzzy rule bases partition the feature space and how easily one system can be translated to another and vice versa. Our complexity metric can also help to choose between different fuzzy classifiers based on what the equivalent crisp partitions look like.
Authors: Modhurita Mitra, Martine G. de Vos, Nicola Cortinovis, Dawa Ometto
Abstract: There has been enormous interest in generative AI since ChatGPT was launched in 2022. However, there are concerns about the accuracy and consistency of the outputs of generative AI. We have carried out an exploratory study on the application of this new technology in research data processing. We identified tasks for which rule-based or traditional machine learning approaches were difficult to apply, and then performed these tasks using generative AI. We demonstrate the feasibility of using the generative AI model Claude 3 Opus in three research projects involving complex data processing tasks: 1) Information extraction: We extract plant species names from historical seedlists (catalogues of seeds) published by botanical gardens. 2) Natural language understanding: We extract certain data points (name of drug, name of health indication, relative effectiveness, cost-effectiveness, etc.) from documents published by Health Technology Assessment organisations in the EU. 3) Text classification: We assign industry codes to projects on the crowdfunding website Kickstarter. We share the lessons we learnt from these use cases: How to determine if generative AI is an appropriate tool for a given data processing task, and if so, how to maximise the accuracy and consistency of the results obtained.
Authors: Xiang Liu, Hau Chan, Minming Li, Xianlong Zeng, Chenchen Fu, Weiwei Wu
Abstract: Federated learning (FL) is a promising approach that allows requesters (\eg, servers) to obtain local training models from workers (e.g., clients). Since workers are typically unwilling to provide training services/models freely and voluntarily, many incentive mechanisms in FL are designed to incentivize participation by offering monetary rewards from requesters. However, existing studies neglect two crucial aspects of real-world FL scenarios. First, workers can possess inherent incompatibility characteristics (e.g., communication channels and data sources), which can lead to degradation of FL efficiency (e.g., low communication efficiency and poor model generalization). Second, the requesters are budgeted, which limits the amount of workers they can hire for their tasks. In this paper, we investigate the scenario in FL where multiple budgeted requesters seek training services from incompatible workers with private training costs. We consider two settings: the cooperative budget setting where requesters cooperate to pool their budgets to improve their overall utility and the non-cooperative budget setting where each requester optimizes their utility within their own budgets. To address efficiency degradation caused by worker incompatibility, we develop novel compatibility-aware incentive mechanisms, CARE-CO and CARE-NO, for both settings to elicit true private costs and determine workers to hire for requesters and their rewards while satisfying requester budget constraints. Our mechanisms guarantee individual rationality, truthfulness, budget feasibility, and approximation performance. We conduct extensive experiments using real-world datasets to show that the proposed mechanisms significantly outperform existing baselines.
Authors: Nikhil Khandalkar, Pavan Yadav, Krishna Shinde, Lokesh B. Ramegowda, Rajarshi Das
Abstract: Recent advancements in Large Language Models (LLMs) have generated growing interest in their structured reasoning capabilities, particularly in tasks involving abstraction and pattern recognition. The Abstraction and Reasoning Corpus (ARC) benchmark plays a crucial role in evaluating these capabilities by testing how well AI models generalize to novel problems. While GPT-4o demonstrates strong performance by solving all ARC tasks under zero-noise conditions, other models like DeepSeek R1 and LLaMA 3.2 fail to solve any, suggesting limitations in their ability to reason beyond simple pattern matching. To explore this gap, we systematically evaluate these models across different noise levels and temperature settings. Our results reveal that the introduction of noise consistently impairs model performance, regardless of architecture. This decline highlights a shared vulnerability: current LLMs, despite showing signs of abstract reasoning, remain highly sensitive to input perturbations. Such fragility raises concerns about their real-world applicability, where noise and uncertainty are common. By comparing how different model architectures respond to these challenges, we offer insights into the structural weaknesses of modern LLMs in reasoning tasks. This work underscores the need for developing more robust and adaptable AI systems capable of handling the ambiguity and variability inherent in real-world scenarios. Our findings aim to guide future research toward enhancing model generalization, robustness, and alignment with human-like cognitive flexibility.
Authors: Isma\"il Baaj
Abstract: In this article, we address the inconsistency of a system of max-min fuzzy relational equations by minimally modifying the matrix governing the system in order to achieve consistency. Our method yields consistent systems that approximate the original inconsistent system in the following sense: the right-hand side vector of each consistent system is that of the inconsistent system, and the coefficients of the matrix governing each consistent system are obtained by modifying, exactly and minimally, the entries of the original matrix that must be corrected to achieve consistency, while leaving all other entries unchanged. To obtain a consistent system that closely approximates the considered inconsistent system, we study the distance (in terms of a norm among $L_1$, $L_2$ or $L_\infty$) between the matrix of the inconsistent system and the set formed by the matrices of consistent systems that use the same right-hand side vector as the inconsistent system. We show that our method allows us to directly compute matrices of consistent systems that use the same right-hand side vector as the inconsistent system whose distance in terms of $L_\infty$ norm to the matrix of the inconsistent system is minimal (the computational costs are higher when using $L_1$ norm or $L_2$ norm). We also give an explicit analytical formula for computing this minimal $L_\infty$ distance. Finally, we translate our results for systems of min-max fuzzy relational equations and present some potential applications.
Authors: Daniele Gorla, Shivam Kumar, Pietro Nicolaus Roselli Lorenzini, Alireza Alipourfaz
Abstract: This paper presents an approach to automating JUnit test generation for Java applications using the Spring Boot framework, leveraging the LLaMA (Large Language Model Architecture) model to enhance the efficiency and accuracy of the testing process. The resulting tool, called CUBETESTERAI, includes a user-friendly web interface and the integration of a CI/CD pipeline using GitLab and Docker. These components streamline the automated test generation process, allowing developers to generate JUnit tests directly from their code snippets with minimal manual intervention. The final implementation executes the LLaMA models through RunPod, an online GPU service, which also enhances the privacy of our tool. Using the advanced natural language processing capabilities of the LLaMA model, CUBETESTERAI is able to generate test cases that provide high code coverage and accurate validation of software functionalities in Java-based Spring Boot applications. Furthermore, it efficiently manages resource-intensive operations and refines the generated tests to address common issues like missing imports and handling of private methods. By comparing CUBETESTERAI with some state-of-the-art tools, we show that our proposal consistently demonstrates competitive and, in many cases, better performance in terms of code coverage in different real-life Java programs.
Authors: Yihong Jin, Ze Yang
Abstract: The rapid expansion of AI inference services in the cloud necessitates a robust scalability solution to manage dynamic workloads and maintain high performance. This study proposes a comprehensive scalability optimization framework for cloud AI inference services, focusing on real-time load balancing and autoscaling strategies. The proposed model is a hybrid approach that combines reinforcement learning for adaptive load distribution and deep neural networks for accurate demand forecasting. This multi-layered approach enables the system to anticipate workload fluctuations and proactively adjust resources, ensuring maximum resource utilisation and minimising latency. Furthermore, the incorporation of a decentralised decision-making process within the model serves to enhance fault tolerance and reduce response time in scaling operations. Experimental results demonstrate that the proposed model enhances load balancing efficiency by 35\ and reduces response delay by 28\, thereby exhibiting a substantial optimization effect in comparison with conventional scalability solutions.
Authors: Haodong Wang, Qihua Zhou, Zicong Hong, Song Guo
Abstract: The mixture of experts (MoE) model is a sparse variant of large language models (LLMs), designed to hold a better balance between intelligent capability and computational overhead. Despite its benefits, MoE is still too expensive to deploy on resource-constrained edge devices, especially with the demands of on-device inference services. Recent research efforts often apply model compression techniques, such as quantization, pruning and merging, to restrict MoE complexity. Unfortunately, due to their predefined static model optimization strategies, they cannot always achieve the desired quality-overhead trade-off when handling multiple requests, finally degrading the on-device quality of service. These limitations motivate us to propose the D$^2$MoE, an algorithm-system co-design framework that matches diverse task requirements by dynamically allocating the most proper bit-width to each expert. Specifically, inspired by the nested structure of matryoshka dolls, we propose the matryoshka weight quantization (MWQ) to progressively compress expert weights in a bit-nested manner and reduce the required runtime memory. On top of it, we further optimize the I/O-computation pipeline and design a heuristic scheduling algorithm following our hottest-expert-bit-first (HEBF) principle, which maximizes the expert parallelism between I/O and computation queue under constrained memory budgets, thus significantly reducing the idle temporal bubbles waiting for the experts to load. Evaluations on real edge devices show that D$^2$MoE improves the overall inference throughput by up to 1.39$\times$ and reduces the peak memory footprint by up to 53% over the latest on-device inference frameworks, while still preserving comparable serving accuracy as its INT8 counterparts.
Authors: Zoi Lygizou, Dimitris Kalles
Abstract: Trust management provides an alternative solution for securing open, dynamic, and distributed multi-agent systems, where conventional cryptographic methods prove to be impractical. However, existing trust models face challenges related to agent mobility, changing behaviors, and the cold start problem. To address these issues we introduced a biologically inspired trust model in which trustees assess their own capabilities and store trust data locally. This design improves mobility support, reduces communication overhead, resists disinformation, and preserves privacy. Despite these advantages, prior evaluations revealed limitations of our model in adapting to provider population changes and continuous performance fluctuations. This study proposes a novel algorithm, incorporating a self-classification mechanism for providers to detect performance drops potentially harmful for the service consumers. Simulation results demonstrate that the new algorithm outperforms its original version and FIRE, a well-known trust and reputation model, particularly in handling dynamic trustee behavior. While FIRE remains competitive under extreme environmental changes, the proposed algorithm demonstrates greater adaptability across various conditions. In contrast to existing trust modeling research, this study conducts a comprehensive evaluation of our model using widely recognized trust model criteria, assessing its resilience against common trust-related attacks while identifying strengths, weaknesses, and potential countermeasures. Finally, several key directions for future research are proposed.
Authors: Yi Xiong, Jinqi Huang, Wenjie Huang, Xuebing Yu, Entong Li, Zhixiong Ning, Jinhua Zhou, Li Zeng, Xin Chen
Abstract: Nowadays, many companies possess various types of AI accelerators, forming heterogeneous clusters. Efficiently leveraging these clusters for high-throughput large language model (LLM) inference services can significantly reduce costs and expedite task processing. However, LLM inference on heterogeneous clusters presents two main challenges. Firstly, different deployment configurations can result in vastly different performance. The number of possible configurations is large, and evaluating the effectiveness of a specific setup is complex. Thus, finding an optimal configuration is not an easy task. Secondly, LLM inference instances within a heterogeneous cluster possess varying processing capacities, leading to different processing speeds for handling inference requests. Evaluating these capacities and designing a request scheduling algorithm that fully maximizes the potential of each instance is challenging. In this paper, we propose a high-throughput inference service system on heterogeneous clusters. First, the deployment configuration is optimized by modeling the resource amount and expected throughput and using the exhaustive search method. Second, a novel mechanism is proposed to schedule requests among instances, which fully considers the different processing capabilities of various instances. Extensive experiments show that the proposed scheduler improves throughput by 122.5% and 33.6% on two heterogeneous clusters, respectively.
Authors: Syeda Tahreem Zahra, Syed Kashif Imdad, Sohail Khan, Sohail Khalid, Nauman Anwar Baig
Abstract: Power transformers play a critical role within the electrical power system, making their health assessment and the prediction of their remaining lifespan paramount for the purpose of ensuring efficient operation and facilitating effective maintenance planning. This paper undertakes a comprehensive examination of existent literature, with a primary focus on both conventional and cutting-edge techniques employed within this domain. The merits and demerits of recent methodologies and techniques are subjected to meticulous scrutiny and explication. Furthermore, this paper expounds upon intelligent fault diagnosis methodologies and delves into the most widely utilized intelligent algorithms for the assessment of transformer conditions. Diverse Artificial Intelligence (AI) approaches, including Artificial Neural Networks (ANN) and Convolutional Neural Network (CNN), Support Vector Machine (SVM), Random Forest (RF), Genetic Algorithm (GA), and Particle Swarm Optimization (PSO), are elucidated offering pragmatic solutions for enhancing the performance of transformer fault diagnosis. The amalgamation of multiple AI methodologies and the exploration of timeseries analysis further contribute to the augmentation of diagnostic precision and the early detection of faults in transformers. By furnishing a comprehensive panorama of AI applications in the field of transformer fault diagnosis, this study lays the groundwork for future research endeavors and the progression of this critical area of study.
Authors: Fei Shang, Haohua Du, Dawei Yan, Panlong Yang, Xiang-Yang Li
Abstract: Due to its ability to work in non-line-of-sight and low-light environments, radio frequency (RF) imaging technology is expected to bring new possibilities for embodied intelligence and multimodal sensing. However, widely used RF devices (such as Wi-Fi) often struggle to provide high-precision electromagnetic measurements and large-scale datasets, hindering the application of RF imaging technology. In this paper, we combine the ideas of PINN to design the RINN network, using physical constraints instead of true value comparison constraints and adapting it with the characteristics of ubiquitous RF signals, allowing the RINN network to achieve RF imaging using only one sample without phase and with amplitude noise. Our numerical evaluation results show that compared with 5 classic algorithms based on phase data for imaging results, RINN's imaging results based on phaseless data are good, with indicators such as RRMSE (0.11) performing similarly well. RINN provides new possibilities for the universal development of radio frequency imaging technology.
Authors: Noa Cohen, Rotem Dror, Itzik Klein
Abstract: Despite the crucial role of inertial measurements in motion tracking and navigation systems, the time-consuming and resource-intensive nature of collecting extensive inertial data has hindered the development of robust machine learning models in this field. In recent years, diffusion models have emerged as a revolutionary class of generative models, reshaping the landscape of artificial data generation. These models surpass generative adversarial networks and other state-of-the-art approaches to complex tasks. In this work, we propose diffusion-driven specific force-generated data for smartphone location recognition. We provide a comprehensive evaluation methodology by comparing synthetic and real recorded specific force data across multiple metrics. Our results demonstrate that our diffusion-based generative model successfully captures the distinctive characteristics of specific force signals across different smartphone placement conditions. Thus, by creating diverse, realistic synthetic data, we can reduce the burden of extensive data collection while providing high-quality training data for machine learning models.
Authors: Meher Boulaabi, Takwa Ben A\"icha Gader, Afef Kacem Echi, Zied Bouraoui
Abstract: Diabetic retinopathy (DR) is a leading cause of blindness worldwide, underscoring the importance of early detection for effective treatment. However, automated DR classification remains challenging due to variations in image quality, class imbalance, and pixel-level similarities that hinder model training. To address these issues, we propose a robust preprocessing pipeline incorporating image cropping, Contrast-Limited Adaptive Histogram Equalization (CLAHE), and targeted data augmentation to improve model generalization and resilience. Our approach leverages the Swin Transformer, which utilizes hierarchical token processing and shifted window attention to efficiently capture fine-grained features while maintaining linear computational complexity. We validate our method on the Aptos and IDRiD datasets for multi-class DR classification, achieving accuracy rates of 89.65% and 97.40%, respectively. These results demonstrate the effectiveness of our model, particularly in detecting early-stage DR, highlighting its potential for improving automated retinal screening in clinical settings.
Authors: Xiao Zhou, Yuze Sun, Jie Wu, Xiaomeng Huang
Abstract: This study introduces ReSA-ConvLSTM, an artificial intelligence (AI) framework for systematic bias correction in numerical weather prediction (NWP). We propose three innovations by integrating dynamic climatological normalization, ConvLSTM with temporal causality constraints, and residual self-attention mechanisms. The model establishes a physics-aware nonlinear mapping between ECMWF forecasts and ERA5 reanalysis data. Using 41 years (1981-2021) of global atmospheric data, the framework reduces systematic biases in 2-m air temperature (T2m), 10-m winds (U10/V10), and sea-level pressure (SLP), achieving up to 20% RMSE reduction over 1-7 day forecasts compared to operational ECMWF outputs. The lightweight architecture (10.6M parameters) enables efficient generalization to multiple variables and downstream applications, reducing retraining time by 85% for cross-variable correction while improving ocean model skill through bias-corrected boundary conditions. The ablation experiments demonstrate that our innovations significantly improve the model's correction performance, suggesting that incorporating variable characteristics into the model helps enhance forecasting skills.
Authors: Donggyun Kim, Chanwoo Kim, Seunghoon Hong
Abstract: While test-time fine-tuning is beneficial in few-shot learning, the need for multiple backpropagation steps can be prohibitively expensive in real-time or low-resource scenarios. To address this limitation, we propose an approach that emulates gradient descent without computing gradients, enabling efficient test-time adaptation. Specifically, we formulate gradient descent as an Euler discretization of an ordinary differential equation (ODE) and train an auxiliary network to predict the task-conditional drift using only the few-shot support set. The adaptation then reduces to a simple numerical integration (e.g., via the Euler method), which requires only a few forward passes of the auxiliary network -- no gradients or forward passes of the target model are needed. In experiments on cross-domain few-shot classification using the Meta-Dataset and CDFSL benchmarks, our method significantly improves out-of-domain performance over the non-fine-tuned baseline while incurring only 6\% of the memory cost and 0.02\% of the computation time of standard fine-tuning, thus establishing a practical middle ground between direct transfer and fully fine-tuned approaches.
Authors: Vuong M. Ngo, Edward Bolger, Stan Goodwin, John O'Sullivan, Dinh Viet Cuong, Mark Roantree
Abstract: Exosomes are small vesicles crucial for cell signaling and disease biomarkers. Due to their complexity, an "omics" approach is preferable to individual biomarkers. While Raman spectroscopy is effective for exosome analysis, it requires high sample concentrations and has limited sensitivity to lipids and proteins. Surface-enhanced Raman spectroscopy helps overcome these challenges. In this study, we leverage Neo4j graph databases to organize 3,045 Raman spectra of exosomes, enhancing data generalization. To further refine spectral analysis, we introduce a novel spectral filtering process that integrates the PageRank Filter with optimal Dimensionality Reduction. This method improves feature selection, resulting in superior classification performance. Specifically, the Extra Trees model, using our spectral processing approach, achieves 0.76 and 0.857 accuracy in classifying hyperglycemic, hypoglycemic, and normal exosome samples based on Raman spectra and surface, respectively, with group 10-fold cross-validation. Our results show that graph-based spectral filtering combined with optimal dimensionality reduction significantly improves classification accuracy by reducing noise while preserving key biomarker signals. This novel framework enhances Raman-based exosome analysis, expanding its potential for biomedical applications, disease diagnostics, and biomarker discovery.
Authors: Alberto Casagrande, Francesco Fabris, Rossano Girometti, Roberto Pagliarini
Abstract: Agreement measures, such as Cohen's kappa or intraclass correlation, gauge the matching between two or more classifiers. They are used in a wide range of contexts from medicine, where they evaluate the effectiveness of medical treatments and clinical trials, to artificial intelligence, where they can quantify the approximation due to the reduction of a classifier. The consistency of different classifiers to a golden standard can be compared simply by using the order induced by their agreement measure with respect to the golden standard itself. Nevertheless, labelling an approach as good or bad exclusively by using the value of an agreement measure requires a scale or a significativity index. Some quality scales have been proposed in the literature for Cohen's kappa, but they are mainly naive, and their boundaries are arbitrary. This work proposes a general approach to evaluate the significativity of any agreement value between two classifiers and introduces two significativity indices: one dealing with finite data sets, the other one handling classification probability distributions. Moreover, this manuscript considers the computational issues of evaluating such indices and identifies some efficient algorithms to evaluate them.
Authors: Usevalad Milasheuski, Luca Barbieri, Sanaz Kianoush, Monica Nicoli, Stefano Savazzi
Abstract: Bayesian Federated Learning (BFL) enables uncertainty quantification and robust adaptation in distributed learning. In contrast to the frequentist approach, it estimates the posterior distribution of a global model, offering insights into model reliability. However, current BFL methods neglect continual learning challenges in dynamic environments where data distributions shift over time. We propose a continual BFL framework applied to human sensing with radar data collected over several days. Using Stochastic Gradient Langevin Dynamics (SGLD), our approach sequentially updates the model, leveraging past posteriors to construct the prior for the new tasks. We assess the accuracy, the expected calibration error (ECE) and the convergence speed of our approach against several baselines. Results highlight the effectiveness of continual Bayesian updates in preserving knowledge and adapting to evolving data.
Authors: Mohit Gupta, Akiko Aizawa, Rajiv Ratn Shah
Abstract: The emergence of large language models (LLMs) has significantly influenced numerous fields, including healthcare, by enhancing the capabilities of automated systems to process and generate human-like text. However, despite their advancements, the reliability and accuracy of LLMs in medical contexts remain critical concerns. Current evaluation methods often lack robustness and fail to provide a comprehensive assessment of LLM performance, leading to potential risks in clinical settings. In this work, we propose Med-CoDE, a specifically designed evaluation framework for medical LLMs to address these challenges. The framework leverages a critique-based approach to quantitatively measure the degree of disagreement between model-generated responses and established medical ground truths. This framework captures both accuracy and reliability in medical settings. The proposed evaluation framework aims to fill the existing gap in LLM assessment by offering a systematic method to evaluate the quality and trustworthiness of medical LLMs. Through extensive experiments and case studies, we illustrate the practicality of our framework in providing a comprehensive and reliable evaluation of medical LLMs.
Authors: Calvin Luo, Zilai Zeng, Yilun Du, Chen Sun
Abstract: Video generative models demonstrate great promise in robotics by serving as visual planners or as policy supervisors. When pretrained on internet-scale data, such video models intimately understand alignment with natural language, and can thus facilitate generalization to novel downstream behavior through text-conditioning. However, they may not be sensitive to the specificities of the particular environment the agent inhabits. On the other hand, training video models on in-domain examples of robotic behavior naturally encodes environment-specific intricacies, but the scale of available demonstrations may not be sufficient to support generalization to unseen tasks via natural language specification. In this work, we investigate different adaptation techniques that integrate in-domain information with large-scale pretrained video models, and explore the extent to which they enable novel text-conditioned generalization for robotic tasks, while also considering their independent data and resource considerations. We successfully demonstrate across robotic environments that adapting powerful video models with small scales of example data can successfully facilitate generalization to novel behaviors. In particular, we present a novel adaptation strategy, termed Inverse Probabilistic Adaptation, that not only consistently achieves strong generalization performance across robotic tasks and settings, but also exhibits robustness to the quality of adaptation data, successfully solving novel tasks even when only suboptimal in-domain demonstrations are available.
Authors: Zhiqiu Lin, Siyuan Cen, Daniel Jiang, Jay Karhade, Hewei Wang, Chancharik Mitra, Tiffany Ling, Yuhan Huang, Sifan Liu, Mingyu Chen, Rushikesh Zawar, Xue Bai, Yilun Du, Chuang Gan, Deva Ramanan
Abstract: We introduce CameraBench, a large-scale dataset and benchmark designed to assess and improve camera motion understanding. CameraBench consists of ~3,000 diverse internet videos, annotated by experts through a rigorous multi-stage quality control process. One of our contributions is a taxonomy of camera motion primitives, designed in collaboration with cinematographers. We find, for example, that some motions like "follow" (or tracking) require understanding scene content like moving subjects. We conduct a large-scale human study to quantify human annotation performance, revealing that domain expertise and tutorial-based training can significantly enhance accuracy. For example, a novice may confuse zoom-in (a change of intrinsics) with translating forward (a change of extrinsics), but can be trained to differentiate the two. Using CameraBench, we evaluate Structure-from-Motion (SfM) and Video-Language Models (VLMs), finding that SfM models struggle to capture semantic primitives that depend on scene content, while VLMs struggle to capture geometric primitives that require precise estimation of trajectories. We then fine-tune a generative VLM on CameraBench to achieve the best of both worlds and showcase its applications, including motion-augmented captioning, video question answering, and video-text retrieval. We hope our taxonomy, benchmark, and tutorials will drive future efforts towards the ultimate goal of understanding camera motions in any video.
Authors: Van-Giang Trinh, Belaid Benhamou, Sylvain Soliman, Fran\c{c}ois Fages
Abstract: Datalog$^\neg$ is a central formalism used in a variety of domains ranging from deductive databases and abstract argumentation frameworks to answer set programming. Its model theory is the finite counterpart of the logical semantics developed for normal logic programs, mainly based on the notions of Clark's completion and two-valued or three-valued canonical models including supported, stable, regular and well-founded models. In this paper we establish a formal link between Datalog$^\neg$ and Boolean network theory, which was initially introduced by Stuart Kaufman and Ren\'e Thomas to reason about gene regulatory networks. We use previous results from Boolean network theory to prove that in the absence of odd cycles in a Datalog$^\neg$ program, the regular models coincide with the stable models, which entails the existence of stable models, and in the absence of even cycles, we show the uniqueness of stable partial models, which entails the uniqueness of regular models. These results on regular models have been claimed by You and Yuan in 1994 for normal logic programs but we show problems in their definition of well-founded stratification and in their proofs that we can fix for negative normal logic programs only. We also give upper bounds on the numbers of stable partial, regular, and stable models of a Datalog$^\neg$ program using the cardinality of a feedback vertex set in its atom dependency graph. Interestingly, our connection to Boolean network theory also points us to the notion of trap spaces for Datalog$^\neg$ programs. We relate the notions of supported or stable trap spaces to the other semantics of Datalog$^\neg$, and show the equivalence between subset-minimal stable trap spaces and regular models.
Authors: Nishath Rajiv Ranasinghe, Shawn M. Jones, Michal Kucer, Ayan Biswas, Daniel O'Malley, Alexander Buschmann Most, Selma Liliane Wanna, Ajay Sreekumar
Abstract: Large Language Models (LLMs) are increasingly being leveraged for generating and translating scientific computer codes by both domain-experts and non-domain experts. Fortran has served as one of the go to programming languages in legacy high-performance computing (HPC) for scientific discoveries. Despite growing adoption, LLM-based code translation of legacy code-bases has not been thoroughly assessed or quantified for its usability. Here, we studied the applicability of LLM-based translation of Fortran to C++ as a step towards building an agentic-workflow using open-weight LLMs on two different computational platforms. We statistically quantified the compilation accuracy of the translated C++ codes, measured the similarity of the LLM translated code to the human translated C++ code, and statistically quantified the output similarity of the Fortran to C++ translation.
Authors: Songyuan Zhang, Oswin So, Mitchell Black, Zachary Serlin, Chuchu Fan
Abstract: Tasks for multi-robot systems often require the robots to collaborate and complete a team goal while maintaining safety. This problem is usually formalized as a constrained Markov decision process (CMDP), which targets minimizing a global cost and bringing the mean of constraint violation below a user-defined threshold. Inspired by real-world robotic applications, we define safety as zero constraint violation. While many safe multi-agent reinforcement learning (MARL) algorithms have been proposed to solve CMDPs, these algorithms suffer from unstable training in this setting. To tackle this, we use the epigraph form for constrained optimization to improve training stability and prove that the centralized epigraph form problem can be solved in a distributed fashion by each agent. This results in a novel centralized training distributed execution MARL algorithm named Def-MARL. Simulation experiments on 8 different tasks across 2 different simulators show that Def-MARL achieves the best overall performance, satisfies safety constraints, and maintains stable training. Real-world hardware experiments on Crazyflie quadcopters demonstrate the ability of Def-MARL to safely coordinate agents to complete complex collaborative tasks compared to other methods.
Authors: Sungjun Han (Trillion Labs), Juyoung Suk (Trillion Labs), Suyeong An (Trillion Labs), Hyungguk Kim (Trillion Labs), Kyuseok Kim (Trillion Labs), Wonsuk Yang (Trillion Labs), Seungtaek Choi (Trillion Labs), Jamin Shin (Trillion Labs)
Abstract: We introduce Trillion-7B, the most token-efficient Korean-centric multilingual LLM available. Our novel Cross-lingual Document Attention (XLDA) mechanism enables highly efficient and effective knowledge transfer from English to target languages like Korean and Japanese. Combined with optimized data mixtures, language-specific filtering, and tailored tokenizer construction, Trillion-7B achieves competitive performance while dedicating only 10\% of its 2T training tokens to multilingual data and requiring just 59.4K H100 GPU hours (\$148K) for full training. Comprehensive evaluations across 27 benchmarks in four languages demonstrate Trillion-7B's robust multilingual performance and exceptional cross-lingual consistency.
Authors: Andrey Fradkin
Abstract: This paper documents three stylized facts about the demand for Large Language Models (LLMs) using data from OpenRouter, a prominent LLM marketplace. First, new models experience rapid initial adoption that stabilizes within weeks. Second, model releases differ substantially in whether they primarily attract new users or substitute demand from competing models. Third, multihoming, using multiple models simultaneously, is common among apps. These findings suggest significant horizontal and vertical differentiation in the LLM market, implying opportunities for providers to maintain demand and pricing power despite rapid technological advances.
Authors: Atin Pothiraj, Elias Stengel-Eskin, Jaemin Cho, Mohit Bansal
Abstract: Recognizing and reasoning about occluded (partially or fully hidden) objects is vital to understanding visual scenes, as occlusions frequently occur in real-world environments and act as obstacles for spatial comprehension. To test models' ability to reason about multiple occluded objects, we introduce a novel task, Counting Amodally for Patterns Through Unseen REgions (CAPTURe), which requires a model to count objects arranged in a pattern by inferring how the pattern continues behind an occluder (an object which blocks parts of the scene). CAPTURe requires both recognizing visual patterns and reasoning, making it a useful testbed for evaluating vision-language models (VLMs) on whether they understand occluded patterns and possess spatial understanding skills. By requiring models to reason about occluded objects, CAPTURe also tests VLMs' ability to form world models that would allow them to fill in missing information. CAPTURe consists of two parts: (1) CAPTURe-real, with manually filtered images of real objects in patterns and (2) CAPTURe-synthetic, a controlled diagnostic with generated patterned images. We evaluate four strong VLMs (GPT-4o, Intern-VL2, Molmo, and Qwen2-VL) on CAPTURe, finding that models struggle to count on both occluded and unoccluded patterns. Crucially, we find that models perform worse with occlusion, suggesting that VLMs are also deficient in inferring unseen spatial relationships: even the strongest VLMs like GPT-4o fail to count with occlusion. In contrast, we find that humans achieve very little error on CAPTURe. We also find that providing auxiliary information of occluded object locations increases performance, underscoring that the model error comes both from an inability to handle occlusion as well as difficulty counting in images.
Authors: Noah Subedar, Taeui Kim, Saathwick Venkataramalingam
Abstract: This paper presents an underlying framework for both automating and accelerating malware classification, more specifically, mapping malicious executables to known Advanced Persistent Threat (APT) groups. The main feature of this analysis is the assembly-level instructions present in executables which are also known as opcodes. The collection of such opcodes on many malicious samples is a lengthy process; hence, open-source reverse engineering tools are used in tandem with scripts that leverage parallel computing to analyze multiple files at once. Traditional and deep learning models are applied to create models capable of classifying malware samples. One-gram and two-gram datasets are constructed and used to train models such as SVM, KNN, and Decision Tree; however, they struggle to provide adequate results without relying on metadata to support n-gram sequences. The computational limitations of such models are overcome with convolutional neural networks (CNNs) and heavily accelerated using graphical compute unit (GPU) resources.
Authors: James Mickens, Sarah Radway, Ravi Netravali
Abstract: As AI models become more embedded in critical sectors like finance, healthcare, and the military, their inscrutable behavior poses ever-greater risks to society. To mitigate this risk, we propose Guillotine, a hypervisor architecture for sandboxing powerful AI models -- models that, by accident or malice, can generate existential threats to humanity. Although Guillotine borrows some well-known virtualization techniques, Guillotine must also introduce fundamentally new isolation mechanisms to handle the unique threat model posed by existential-risk AIs. For example, a rogue AI may try to introspect upon hypervisor software or the underlying hardware substrate to enable later subversion of that control plane; thus, a Guillotine hypervisor requires careful co-design of the hypervisor software and the CPUs, RAM, NIC, and storage devices that support the hypervisor software, to thwart side channel leakage and more generally eliminate mechanisms for AI to exploit reflection-based vulnerabilities. Beyond such isolation at the software, network, and microarchitectural layers, a Guillotine hypervisor must also provide physical fail-safes more commonly associated with nuclear power plants, avionic platforms, and other types of mission critical systems. Physical fail-safes, e.g., involving electromechanical disconnection of network cables, or the flooding of a datacenter which holds a rogue AI, provide defense in depth if software, network, and microarchitectural isolation is compromised and a rogue AI must be temporarily shut down or permanently destroyed.
Authors: Wuchen Li
Abstract: We define a class of divergences to measure differences between probability density functions in one-dimensional sample space. The construction is based on the convex function with the Jacobi operator of mapping function that pushforwards one density to the other. We call these information measures {\em transport $f$-divergences}. We present several properties of transport $f$-divergences, including invariances, convexities, variational formulations, and Taylor expansions in terms of mapping functions. Examples of transport $f$-divergences in generative models are provided.
Authors: Qiyao Wang, Guhong Chen, Hongbo Wang, Huaren Liu, Minghui Zhu, Zhifei Qin, Linwei Li, Yilin Yue, Shiqiang Wang, Jiayan Li, Yihang Wu, Ziqiang Liu, Longze Chen, Run Luo, Liyang Fan, Jiaming Li, Lei Zhang, Kan Xu, Hongfei Lin, Hamid Alinejad-Rokny, Shiwen Ni, Yuan Lin, Min Yang
Abstract: Intellectual Property (IP) is a unique domain that integrates technical and legal knowledge, making it inherently complex and knowledge-intensive. As large language models (LLMs) continue to advance, they show great potential for processing IP tasks, enabling more efficient analysis, understanding, and generation of IP-related content. However, existing datasets and benchmarks either focus narrowly on patents or cover limited aspects of the IP field, lacking alignment with real-world scenarios. To bridge this gap, we introduce the first comprehensive IP task taxonomy and a large, diverse bilingual benchmark, IPBench, covering 8 IP mechanisms and 20 tasks. This benchmark is designed to evaluate LLMs in real-world intellectual property applications, encompassing both understanding and generation. We benchmark 16 LLMs, ranging from general-purpose to domain-specific models, and find that even the best-performing model achieves only 75.8% accuracy, revealing substantial room for improvement. Notably, open-source IP and law-oriented models lag behind closed-source general-purpose models. We publicly release all data and code of IPBench and will continue to update it with additional IP-related tasks to better reflect real-world challenges in the intellectual property domain.
Authors: Jayachandu Bandlamudi, Ritwik Chaudhuri, Neelamadhav Gantayat, Kushal Mukherjee, Prerna Agarwal, Renuka Sindhgatta, Sameep Mehta
Abstract: Large Language Models (LLMs) are enabling autonomous agents to perform complex workflows using external tools or functions, often provided via REST APIs in enterprise systems. However, directly utilizing these APIs as tools poses challenges due to their complex input schemas, elaborate responses, and often ambiguous documentation. Current benchmarks for tool testing do not adequately address these complexities, leading to a critical gap in evaluating API readiness for agent-driven automation. In this work, we present a novel testing framework aimed at evaluating and enhancing the readiness of REST APIs to function as tools for LLM-based agents. Our framework transforms apis as tools, generates comprehensive test cases for the APIs, translates tests cases into natural language instructions suitable for agents, enriches tool definitions and evaluates the agent's ability t correctly invoke the API and process its inputs and responses. To provide actionable insights, we analyze the outcomes of 750 test cases, presenting a detailed taxonomy of errors, including input misinterpretation, output handling inconsistencies, and schema mismatches. Additionally, we classify these test cases to streamline debugging and refinement of tool integrations. This work offers a foundational step toward enabling enterprise APIs as tools, improving their usability in agent-based applications.
Authors: Anjali Khurana, Xiaotian Su, April Yi Wang, Parmit K Chilana
Abstract: Large Language Model (LLM)-based in-application assistants, or copilots, can automate software tasks, but users often prefer learning by doing, raising questions about the optimal level of automation for an effective user experience. We investigated two automation paradigms by designing and implementing a fully automated copilot (AutoCopilot) and a semi-automated copilot (GuidedCopilot) that automates trivial steps while offering step-by-step visual guidance. In a user study (N=20) across data analysis and visual design tasks, GuidedCopilot outperformed AutoCopilot in user control, software utility, and learnability, especially for exploratory and creative tasks, while AutoCopilot saved time for simpler visual tasks. A follow-up design exploration (N=10) enhanced GuidedCopilot with task-and state-aware features, including in-context preview clips and adaptive instructions. Our findings highlight the critical role of user control and tailored guidance in designing the next generation of copilots that enhance productivity, support diverse skill levels, and foster deeper software engagement.
Authors: Musfiqur Rahman, SayedHassan Khatoonabadi, Emad Shihab
Abstract: Recent advancements in large language models (LLMs) have demonstrated promising capabilities in code generation tasks. However, most existing benchmarks focus on isolated functions and fail to capture the complexity of real-world, class-level software structures. To address this gap, we introduce a large-scale, Python class-level dataset curated from $13{,}174$ real-world open-source projects. The dataset contains over 842,000 class skeletons, each including class and method signatures, along with associated docstrings when available. We preserve structural and contextual dependencies critical to realistic software development scenarios and enrich the dataset with static code metrics to support downstream analysis. To evaluate the usefulness of this dataset, we use extracted class skeletons as prompts for GPT-4 to generate full class implementations. Results show that the LLM-generated classes exhibit strong lexical and structural similarity to human-written counterparts, with average ROUGE@L, BLEU, and TSED scores of 0.80, 0.59, and 0.73, respectively. These findings confirm that well-structured prompts derived from real-world class skeletons significantly enhance LLM performance in class-level code generation. This dataset offers a valuable resource for benchmarking, training, and improving LLMs in realistic software engineering contexts.
Authors: Kun Wang, Guibin Zhang, Zhenhong Zhou, Jiahao Wu, Miao Yu, Shiqian Zhao, Chenlong Yin, Jinhu Fu, Yibo Yan, Hanjun Luo, Liang Lin, Zhihao Xu, Haolang Lu, Xinye Cao, Xinyun Zhou, Weifei Jin, Fanci Meng, Junyuan Mao, Hao Wu, Minghe Wang, Fan Zhang, Junfeng Fang, Chengwei Liu, Yifan Zhang, Qiankun Li, Chongye Guo, Yalan Qin, Yi Ding, Donghai Hong, Jiaming Ji, Xinfeng Li, Yifan Jiang, Dongxia Wang, Yihao Huang, Yufei Guo, Jen-tse Huang, Yanwei Yue, Wenke Huang, Guancheng Wan, Tianlin Li, Lei Bai, Jie Zhang, Qing Guo, Jingyi Wang, Tianlong Chen, Joey Tianyi Zhou, Xiaojun Jia, Weisong Sun, Cong Wu, Jing Chen, Xuming Hu, Yiming Li, Xiao Wang, Ningyu Zhang, Luu Anh Tuan, Guowen Xu, Tianwei Zhang, Xingjun Ma, Xiang Wang, Bo An, Jun Sun, Mohit Bansal, Shirui Pan, Yuval Elovici, Bhavya Kailkhura, Bo Li, Yaodong Yang, Hongwei Li, Wenyuan Xu, Yizhou Sun, Wei Wang, Qing Li, Ke Tang, Yu-Gang Jiang, Felix Juefei-Xu, Hui Xiong, Xiaofeng Wang, Shuicheng Yan, Dacheng Tao, Philip S. Yu, Qingsong Wen, Yang Liu
Abstract: The remarkable success of Large Language Models (LLMs) has illuminated a promising pathway toward achieving Artificial General Intelligence for both academic and industrial communities, owing to their unprecedented performance across various applications. As LLMs continue to gain prominence in both research and commercial domains, their security and safety implications have become a growing concern, not only for researchers and corporations but also for every nation. Currently, existing surveys on LLM safety primarily focus on specific stages of the LLM lifecycle, e.g., deployment phase or fine-tuning phase, lacking a comprehensive understanding of the entire "lifechain" of LLMs. To address this gap, this paper introduces, for the first time, the concept of "full-stack" safety to systematically consider safety issues throughout the entire process of LLM training, deployment, and eventual commercialization. Compared to the off-the-shelf LLM safety surveys, our work demonstrates several distinctive advantages: (I) Comprehensive Perspective. We define the complete LLM lifecycle as encompassing data preparation, pre-training, post-training, deployment and final commercialization. To our knowledge, this represents the first safety survey to encompass the entire lifecycle of LLMs. (II) Extensive Literature Support. Our research is grounded in an exhaustive review of over 800+ papers, ensuring comprehensive coverage and systematic organization of security issues within a more holistic understanding. (III) Unique Insights. Through systematic literature analysis, we have developed reliable roadmaps and perspectives for each chapter. Our work identifies promising research directions, including safety in data generation, alignment techniques, model editing, and LLM-based agent systems. These insights provide valuable guidance for researchers pursuing future work in this field.
Authors: Zimo Yan, Jie Zhang, Zheng Xie, Chang Liu, Yizhen Liu, Yiping Song
Abstract: Molecular generation plays an important role in drug discovery and materials science, especially in data-scarce scenarios where traditional generative models often struggle to achieve satisfactory conditional generalization. To address this challenge, we propose MetaMolGen, a first-order meta-learning-based molecular generator designed for few-shot and property-conditioned molecular generation. MetaMolGen standardizes the distribution of graph motifs by mapping them to a normalized latent space, and employs a lightweight autoregressive sequence model to generate SMILES sequences that faithfully reflect the underlying molecular structure. In addition, it supports conditional generation of molecules with target properties through a learnable property projector integrated into the generative process.Experimental results demonstrate that MetaMolGen consistently generates valid and diverse SMILES sequences under low-data regimes, outperforming conventional baselines. This highlights its advantage in fast adaptation and efficient conditional generation for practical molecular design.
Authors: Pavan Yadav, Nikhil Khandalkar, Krishna Shinde, Lokesh B. Ramegowda, Rajarshi Das
Abstract: Language models have made significant progress in generating coherent text and predicting next tokens based on input prompts. This study compares the next-token prediction performance of two well-known models: OpenAI's GPT-2 and Meta's Llama-2-7b-chat-hf on Theory of Mind (ToM) tasks. To evaluate their capabilities, we built a dataset from 10 short stories sourced from the Explore ToM Dataset. We enhanced these stories by programmatically inserting additional sentences (infills) using GPT-4, creating variations that introduce different levels of contextual complexity. This setup enables analysis of how increasing context affects model performance. We tested both models under four temperature settings (0.01, 0.5, 1.0, 2.0) and evaluated their ability to predict the next token across three reasoning levels. Zero-order reasoning involves tracking the state, either current (ground truth) or past (memory). First-order reasoning concerns understanding another's mental state (e.g., "Does Anne know the apple is salted?"). Second-order reasoning adds recursion (e.g., "Does Anne think that Charles knows the apple is salted?"). Our results show that adding more infill sentences slightly reduces prediction accuracy, as added context increases complexity and ambiguity. Llama-2 consistently outperforms GPT-2 in prediction accuracy, especially at lower temperatures, demonstrating greater confidence in selecting the most probable token. As reasoning complexity rises, model responses diverge more. Notably, GPT-2 and Llama-2 display greater variability in predictions during first- and second-order reasoning tasks. These findings illustrate how model architecture, temperature, and contextual complexity influence next-token prediction, contributing to a better understanding of the strengths and limitations of current language models.
Authors: Peizheng Liu, Hitoshi Iba
Abstract: Transformer-based architectures have recently propelled advances in sequence modeling across domains, but their application to the hydrophobic-hydrophilic (H-P) model for protein folding remains relatively unexplored. In this work, we adapt a Deep Q-Network (DQN) integrated with attention mechanisms (Transformers) to address the 3D H-P protein folding problem. Our system formulates folding decisions as a self-avoiding walk in a reinforced environment, and employs a specialized reward function based on favorable hydrophobic interactions. To improve performance, the method incorporates validity check including symmetry-breaking constraints, dueling and double Q-learning, and prioritized replay to focus learning on critical transitions. Experimental evaluations on standard benchmark sequences demonstrate that our approach achieves several known best solutions for shorter sequences, and obtains near-optimal results for longer chains. This study underscores the promise of attention-based reinforcement learning for protein folding, and created a prototype of Transformer-based Q-network structure for 3-dimensional lattice models.
Authors: Farnaz Behrang, Zhizhou Zhang, Georgian-Vlad Saioc, Peng Liu, Milind Chabbi
Abstract: Data races are a prevalent class of concurrency bugs in shared-memory parallel programs, posing significant challenges to software reliability and reproducibility. While there is an extensive body of research on detecting data races and a wealth of practical detection tools across various programming languages, considerably less effort has been directed toward automatically fixing data races at an industrial scale. In large codebases, data races are continuously introduced and exhibit myriad patterns, making automated fixing particularly challenging. In this paper, we tackle the problem of automatically fixing data races at an industrial scale. We present Dr.Fix, a tool that combines large language models (LLMs) with program analysis to generate fixes for data races in real-world settings, effectively addressing a broad spectrum of racy patterns in complex code contexts. Implemented for Go--the programming language widely used in modern microservice architectures where concurrency is pervasive and data races are common--Dr.Fix seamlessly integrates into existing development workflows. We detail the design of Dr.Fix and examine how individual design choices influence the quality of the fixes produced. Over the past 18 months, Dr.Fix has been integrated into developer workflows at Uber demonstrating its practical utility. During this period, Dr.Fix produced patches for 224 (55%) from a corpus of 404 data races spanning various categories; 193 of these patches (86%) were accepted by more than a hundred developers via code reviews and integrated into the codebase.
Authors: Hongtao Wang, Taiyan Zhang, Renchi Yang, Jianliang Xu
Abstract: Text clustering aims to automatically partition a collection of text documents into distinct clusters based on linguistic features. In the literature, this task is usually framed as metric clustering based on text embeddings from pre-trained encoders or a graph clustering problem upon pairwise similarities from an oracle, e.g., a large ML model. Recently, large language models (LLMs) bring significant advancement in this field by offering contextualized text embeddings and highly accurate similarity scores, but meanwhile, present grand challenges to cope with substantial computational and/or financial overhead caused by numerous API-based queries or inference calls to the models. In response, this paper proposes TECL, a cost-effective framework that taps into the feedback from LLMs for accurate text clustering within a limited budget of queries to LLMs. Under the hood, TECL adopts our EdgeLLM or TriangleLLM to construct must-link/cannot-link constraints for text pairs, and further leverages such constraints as supervision signals input to our weighted constrained clustering approach to generate clusters. Particularly, EdgeLLM (resp. TriangleLLM) enables the identification of informative text pairs (resp. triplets) for querying LLMs via well-thought-out greedy algorithms and accurate extraction of pairwise constraints through carefully-crafted prompting techniques. Our experiments on multiple benchmark datasets exhibit that TECL consistently and considerably outperforms existing solutions in unsupervised text clustering under the same query cost for LLMs.
Authors: Md Abdul Baset Sarker, Art Nguyen, Sigmond Kukla, Kevin Fite, Masudul H. Imtiaz
Abstract: This paper introduces a novel AI vision-enabled pediatric prosthetic hand designed to assist children aged 10-12 with upper limb disabilities. The prosthesis features an anthropomorphic appearance, multi-articulating functionality, and a lightweight design that mimics a natural hand, making it both accessible and affordable for low-income families. Using 3D printing technology and integrating advanced machine vision, sensing, and embedded computing, the prosthetic hand offers a low-cost, customizable solution that addresses the limitations of current myoelectric prostheses. A micro camera is interfaced with a low-power FPGA for real-time object detection and assists with precise grasping. The onboard DL-based object detection and grasp classification models achieved accuracies of 96% and 100% respectively. In the force prediction, the mean absolute error was found to be 0.018. The features of the proposed prosthetic hand can thus be summarized as: a) a wrist-mounted micro camera for artificial sensing, enabling a wide range of hand-based tasks; b) real-time object detection and distance estimation for precise grasping; and c) ultra-low-power operation that delivers high performance within constrained power and resource limits.
Authors: Anjiang Wei, Huanmi Tan, Tarun Suresh, Daniel Mendoza, Thiago S. F. X. Teixeira, Ke Wang, Caroline Trippel, Alex Aiken
Abstract: Recent advances in Large Language Models (LLMs) have sparked growing interest in applying them to Electronic Design Automation (EDA) tasks, particularly Register Transfer Level (RTL) code generation. While several RTL datasets have been introduced, most focus on syntactic validity rather than functional validation with tests, leading to training examples that compile but may not implement the intended behavior. We present VERICODER, a model for RTL code generation fine-tuned on a dataset validated for functional correctness. This fine-tuning dataset is constructed using a novel methodology that combines unit test generation with feedback-directed refinement. Given a natural language specification and an initial RTL design, we prompt a teacher model (GPT-4o-mini) to generate unit tests and iteratively revise the RTL design based on its simulation results using the generated tests. If necessary, the teacher model also updates the tests to ensure they comply with the natural language specification. As a result of this process, every example in our dataset is functionally validated, consisting of a natural language description, an RTL implementation, and passing tests. Fine-tuned on this dataset of over 125,000 examples, VERICODER achieves state-of-the-art metrics in functional correctness on VerilogEval and RTLLM, with relative gains of up to 71.7% and 27.4% respectively. An ablation study further shows that models trained on our functionally validated dataset outperform those trained on functionally non-validated datasets, underscoring the importance of high-quality datasets in RTL code generation.
Authors: Ju Yeon Kang, Ji Won Yoon, Semin Kim, Min Hyun Han, Nam Soo Kim
Abstract: Recently, fake audio detection has gained significant attention, as advancements in speech synthesis and voice conversion have increased the vulnerability of automatic speaker verification (ASV) systems to spoofing attacks. A key challenge in this task is generalizing models to detect unseen, out-of-distribution (OOD) attacks. Although existing approaches have shown promising results, they inherently suffer from overconfidence issues due to the usage of softmax for classification, which can produce unreliable predictions when encountering unpredictable spoofing attempts. To deal with this limitation, we propose a novel framework called fake audio detection with evidential learning (FADEL). By modeling class probabilities with a Dirichlet distribution, FADEL incorporates model uncertainty into its predictions, thereby leading to more robust performance in OOD scenarios. Experimental results on the ASVspoof2019 Logical Access (LA) and ASVspoof2021 LA datasets indicate that the proposed method significantly improves the performance of baseline models. Furthermore, we demonstrate the validity of uncertainty estimation by analyzing a strong correlation between average uncertainty and equal error rate (EER) across different spoofing algorithms.
Authors: Yannic Neuhaus, Matthias Hein
Abstract: Since data annotation is costly, benchmark datasets often incorporate labels from established image datasets. In this work, we assess the impact of label errors in MSCOCO on the frequently used object hallucination benchmark POPE. We re-annotate the benchmark images and identify an imbalance in annotation errors across different subsets. Evaluating multiple models on the revised labels, which we denote as RePOPE, we observe notable shifts in model rankings, highlighting the impact of label quality. Code and data are available at https://github.com/YanNeu/RePOPE .
Authors: Yiannis Papageorgiou, Yannis Thomas, Alexios Filippakopoulos, Ramin Khalili, Iordanis Koutsopoulos
Abstract: Federated learning (FL) operates based on model exchanges between the server and the clients, and it suffers from significant client-side computation and communication burden. Split federated learning (SFL) arises a promising solution by splitting the model into two parts, that are trained sequentially: the clients train the first part of the model (client-side model) and transmit it to the server that trains the second (server-side model). Existing SFL schemes though still exhibit long training delays and significant communication overhead, especially when clients of different computing capability participate. Thus, we propose Collaborative-Split Federated Learning~(C-SFL), a novel scheme that splits the model into three parts, namely the model parts trained at the computationally weak clients, the ones trained at the computationally strong clients, and the ones at the server. Unlike existing works, C-SFL enables parallel training and aggregation of model's parts at the clients and at the server, resulting in reduced training delays and commmunication overhead while improving the model's accuracy. Experiments verify the multiple gains of C-SFL against the existing schemes.
Authors: Seung Gyu Jeong, Sung Woo Nam, Seong Kwan Jung, Seong-Eun Kim
Abstract: Respiratory auscultation is crucial for early detection of pediatric pneumonia, a condition that can quickly worsen without timely intervention. In areas with limited physician access, effective auscultation is challenging. We present a smartphone-based system that leverages built-in microphones and advanced deep learning algorithms to detect abnormal respiratory sounds indicative of pneumonia risk. Our end-to-end deep learning framework employs domain generalization to integrate a large electronic stethoscope dataset with a smaller smartphone-derived dataset, enabling robust feature learning for accurate respiratory assessments without expensive equipment. The accompanying mobile application guides caregivers in collecting high-quality lung sound samples and provides immediate feedback on potential pneumonia risks. User studies show strong classification performance and high acceptance, demonstrating the system's ability to facilitate proactive interventions and reduce preventable childhood pneumonia deaths. By seamlessly integrating into ubiquitous smartphones, this approach offers a promising avenue for more equitable and comprehensive remote pediatric care.
Authors: Tobias Demmler, Lennart Hartung, Andreas Tamke, Thao Dang, Alexander Hegai, Karsten Haug, Lars Mikelsons
Abstract: In autonomous driving, accurately predicting the movements of other traffic participants is crucial, as it significantly influences a vehicle's planning processes. Modern trajectory prediction models strive to interpret complex patterns and dependencies from agent and map data. The Motion Transformer (MTR) architecture and subsequent work define the most accurate methods in common benchmarks such as the Waymo Open Motion Benchmark. The MTR model employs pre-generated static intention points as initial goal points for trajectory prediction. However, the static nature of these points frequently leads to misalignment with map data in specific traffic scenarios, resulting in unfeasible or unrealistic goal points. Our research addresses this limitation by integrating scene-specific dynamic intention points into the MTR model. This adaptation of the MTR model was trained and evaluated on the Waymo Open Motion Dataset. Our findings demonstrate that incorporating dynamic intention points has a significant positive impact on trajectory prediction accuracy, especially for predictions over long time horizons. Furthermore, we analyze the impact on ground truth trajectories which are not compliant with the map data or are illegal maneuvers.
Authors: Cong Liu, Sharvaree Vadgama, David Ruhe, Erik Bekkers, Patrick Forr\`e
Abstract: This paper explores leveraging the Clifford algebra's expressive power for $\E(n)$-equivariant diffusion models. We utilize the geometric products between Clifford multivectors and the rich geometric information encoded in Clifford subspaces in \emph{Clifford Diffusion Models} (CDMs). We extend the diffusion process beyond just Clifford one-vectors to incorporate all higher-grade multivector subspaces. The data is embedded in grade-$k$ subspaces, allowing us to apply latent diffusion across complete multivectors. This enables CDMs to capture the joint distribution across different subspaces of the algebra, incorporating richer geometric information through higher-order features. We provide empirical results for unconditional molecular generation on the QM9 dataset, showing that CDMs provide a promising avenue for generative modeling.
Authors: Aaron J. Gutknecht, Fernando E. Rosas, David A. Ehrlich, Abdullah Makkeh, Pedro A. M. Mediano, Michael Wibral
Abstract: Distributed systems, such as biological and artificial neural networks, process information via complex interactions engaging multiple subsystems, resulting in high-order patterns with distinct properties across scales. Investigating how these systems process information remains challenging due to difficulties in defining appropriate multivariate metrics and ensuring their scalability to large systems. To address these challenges, we introduce a novel framework based on what we call "Shannon invariants" -- quantities that capture essential properties of high-order information processing in a way that depends only on the definition of entropy and can be efficiently calculated for large systems. Our theoretical results demonstrate how Shannon invariants can be used to resolve long-standing ambiguities regarding the interpretation of widely used multivariate information-theoretic measures. Moreover, our practical results reveal distinctive information-processing signatures of various deep learning architectures across layers, which lead to new insights into how these systems process information and how this evolves during training. Overall, our framework resolves fundamental limitations in analyzing high-order phenomena and offers broad opportunities for theoretical developments and empirical analyses.
Authors: Ruizhe Li, Chiwei Zhu, Benfeng Xu, Xiaorui Wang, Zhendong Mao
Abstract: Creative writing is a key capability of Large Language Models (LLMs), with potential applications in literature, storytelling, and various creative domains. However, evaluating the creativity of machine-generated texts remains a significant challenge, as existing methods either rely on costly manual annotations or fail to align closely with human assessments. In this paper, we propose an effective automated evaluation method based on the Torrance Test of Creative Writing (TTCW), which evaluates creativity as product. Our method employs a reference-based Likert-style approach, scoring generated creative texts relative to high-quality reference texts across various tests. Experimental results demonstrate that our method significantly improves the alignment between LLM evaluations and human assessments, achieving a pairwise accuracy of 0.75 (+15\%).
Authors: Valeria Lerman, Yaniv Dover
Abstract: As large language models (LLMs) and LLM-based agents increasingly interact with humans in decision-making contexts, understanding the trust dynamics between humans and AI agents becomes a central concern. While considerable literature studies how humans trust AI agents, it is much less understood how LLM-based agents develop effective trust in humans. LLM-based agents likely rely on some sort of implicit effective trust in trust-related contexts (e.g., evaluating individual loan applications) to assist and affect decision making. Using established behavioral theories, we develop an approach that studies whether LLMs trust depends on the three major trustworthiness dimensions: competence, benevolence and integrity of the human subject. We also study how demographic variables affect effective trust. Across 43,200 simulated experiments, for five popular language models, across five different scenarios we find that LLM trust development shows an overall similarity to human trust development. We find that in most, but not all cases, LLM trust is strongly predicted by trustworthiness, and in some cases also biased by age, religion and gender, especially in financial scenarios. This is particularly true for scenarios common in the literature and for newer models. While the overall patterns align with human-like mechanisms of effective trust formation, different models exhibit variation in how they estimate trust; in some cases, trustworthiness and demographic factors are weak predictors of effective trust. These findings call for a better understanding of AI-to-human trust dynamics and monitoring of biases and trust development patterns to prevent unintended and potentially harmful outcomes in trust-sensitive applications of AI.
Authors: Ning Wang, Bingkun Yao, Jie Zhou, Yuchen Hu, Xi Wang, Nan Guan, Zhe Jiang
Abstract: Large language models (LLMs) have shown strong performance in Verilog generation from natural language description. However, ensuring the functional correctness of the generated code remains a significant challenge. This paper introduces a method that integrates verification insights from testbench into the training of Verilog generation LLMs, aligning the training with the fundamental goal of hardware design: functional correctness. The main obstacle in using LLMs for Verilog code generation is the lack of sufficient functional verification data, particularly testbenches paired with design specifications and code. To address this problem, we introduce an automatic testbench generation pipeline that decomposes the process and uses feedback from the Verilog compiler simulator (VCS) to reduce hallucination and ensure correctness. We then use the testbench to evaluate the generated codes and collect them for further training, where verification insights are introduced. Our method applies reinforcement learning (RL), specifically direct preference optimization (DPO), to align Verilog code generation with functional correctness by training preference pairs based on testbench outcomes. In evaluations on VerilogEval-Machine, VerilogEval-Human, RTLLM v1.1, RTLLM v2, and VerilogEval v2, our approach consistently outperforms state-of-the-art baselines in generating functionally correct Verilog code. We open source all training code, data, and models at https://anonymous.4open.science/r/VeriPrefer-E88B.
Authors: Kai Luo, Juan Tang, Mingchao Cai, Xiaoqing Zeng, Manqi Xie, Ming Yan
Abstract: Kolmogorov-Arnold Networks (KANs) have emerged as a promising alternative to Multi-Layer Perceptrons (MLPs) due to their superior function-fitting abilities in data-driven modeling. In this paper, we propose a novel framework, DAE-KAN, for solving high-index differential-algebraic equations (DAEs) by integrating KANs with Physics-Informed Neural Networks (PINNs). This framework not only preserves the ability of traditional PINNs to model complex systems governed by physical laws but also enhances their performance by leveraging the function-fitting strengths of KANs. Numerical experiments demonstrate that for DAE systems ranging from index-1 to index-3, DAE-KAN reduces the absolute errors of both differential and algebraic variables by 1 to 2 orders of magnitude compared to traditional PINNs. To assess the effectiveness of this approach, we analyze the drift-off error and find that both PINNs and DAE-KAN outperform classical numerical methods in controlling this phenomenon. Our results highlight the potential of neural network methods, particularly DAE-KAN, in solving high-index DAEs with substantial computational accuracy and generalization, offering a promising solution for challenging partial differential-algebraic equations.
Authors: Xuchuang Wang, Qirun Zeng, Jinhang Zuo, Xutong Liu, Mohammad Hajiesmaili, John C. S. Lui, Adam Wierman
Abstract: This paper investigates the fusion of absolute (reward) and relative (dueling) feedback in stochastic bandits, where both feedback types are gathered in each decision round. We derive a regret lower bound, demonstrating that an efficient algorithm may incur only the smaller among the reward and dueling-based regret for each individual arm. We propose two fusion approaches: (1) a simple elimination fusion algorithm that leverages both feedback types to explore all arms and unifies collected information by sharing a common candidate arm set, and (2) a decomposition fusion algorithm that selects the more effective feedback to explore the corresponding arms and randomly assigns one feedback type for exploration and the other for exploitation in each round. The elimination fusion experiences a suboptimal multiplicative term of the number of arms in regret due to the intrinsic suboptimality of dueling elimination. In contrast, the decomposition fusion achieves regret matching the lower bound up to a constant under a common assumption. Extensive experiments confirm the efficacy of our algorithms and theoretical results.
Authors: Songyan Xie, Jinghang Wen, Encheng Su, Qiucheng Yu
Abstract: Near-infrared (NIR) face recognition systems, which can operate effectively in low-light conditions or in the presence of makeup, exhibit vulnerabilities when subjected to physical adversarial attacks. To further demonstrate the potential risks in real-world applications, we design a novel, stealthy, and practical adversarial patch to attack NIR face recognition systems in a black-box setting. We achieved this by utilizing human-imperceptible infrared-absorbing ink to generate multiple patches with digitally optimized shapes and positions for infrared images. To address the optimization mismatch between digital and real-world NIR imaging, we develop a light reflection model for human skin to minimize pixel-level discrepancies by simulating NIR light reflection. Compared to state-of-the-art (SOTA) physical attacks on NIR face recognition systems, the experimental results show that our method improves the attack success rate in both digital and physical domains, particularly maintaining effectiveness across various face postures. Notably, the proposed approach outperforms SOTA methods, achieving an average attack success rate of 82.46% in the physical domain across different models, compared to 64.18% for existing methods. The artifact is available at https://anonymous.4open.science/r/Human-imperceptible-adversarial-patch-0703/.
URLs: https://anonymous.4open.science/r/Human-imperceptible-adversarial-patch-0703/.
Authors: Xuyang Zhong, Haochen Luo, Chen Liu
Abstract: Existing machine unlearning (MU) approaches exhibit significant sensitivity to hyperparameters, requiring meticulous tuning that limits practical deployment. In this work, we first empirically demonstrate the instability and suboptimal performance of existing popular MU methods when deployed in different scenarios. To address this issue, we propose Dual Optimizer (DualOptim), which incorporates adaptive learning rate and decoupled momentum factors. Empirical and theoretical evidence demonstrates that DualOptim contributes to effective and stable unlearning. Through extensive experiments, we show that DualOptim can significantly boost MU efficacy and stability across diverse tasks, including image classification, image generation, and large language models, making it a versatile approach to empower existing MU algorithms.
Authors: Lotfi Abdelkrim Mecharbat, Ibrahim Elmakky, Martin Takac, Mohammed Yaqub
Abstract: Deep learning (DL) has achieved remarkable progress in the field of medical imaging. However, adapting DL models to medical tasks remains a significant challenge, primarily due to two key factors: (1) architecture selection, as different tasks necessitate specialized model designs, and (2) weight initialization, which directly impacts the convergence speed and final performance of the models. Although transfer learning from ImageNet is a widely adopted strategy, its effectiveness is constrained by the substantial differences between natural and medical images. To address these challenges, we introduce Medical Neural Network Search (MedNNS), the first Neural Network Search framework for medical imaging applications. MedNNS jointly optimizes architecture selection and weight initialization by constructing a meta-space that encodes datasets and models based on how well they perform together. We build this space using a Supernetwork-based approach, expanding the model zoo size by 51x times over previous state-of-the-art (SOTA) methods. Moreover, we introduce rank loss and Fr\'echet Inception Distance (FID) loss into the construction of the space to capture inter-model and inter-dataset relationships, thereby achieving more accurate alignment in the meta-space. Experimental results across multiple datasets demonstrate that MedNNS significantly outperforms both ImageNet pre-trained DL models and SOTA Neural Architecture Search (NAS) methods, achieving an average accuracy improvement of 1.7% across datasets while converging substantially faster. The code and the processed meta-space is available at https://github.com/BioMedIA-MBZUAI/MedNNS.
Authors: Qizhen Wu Lei Chen, Kexin Liu, Jinhu L\"u
Abstract: In swarm robotics, confrontation scenarios, including strategic confrontations, require efficient decision-making that integrates discrete commands and continuous actions. Traditional task and motion planning methods separate decision-making into two layers, but their unidirectional structure fails to capture the interdependence between these layers, limiting adaptability in dynamic environments. Here, we propose a novel bidirectional approach based on hierarchical reinforcement learning, enabling dynamic interaction between the layers. This method effectively maps commands to task allocation and actions to path planning, while leveraging cross-training techniques to enhance learning across the hierarchical framework. Furthermore, we introduce a trajectory prediction model that bridges abstract task representations with actionable planning goals. In our experiments, it achieves over 80\% in confrontation win rate and under 0.01 seconds in decision time, outperforming existing approaches. Demonstrations through large-scale tests and real-world robot experiments further emphasize the generalization capabilities and practical applicability of our method.
Authors: Farida Mohsen, Samir Belhaouari, Zubair Shah
Abstract: Diabetic retinopathy is a serious ocular complication that poses a significant threat to patients' vision and overall health. Early detection and accurate grading are essential to prevent vision loss. Current automatic grading methods rely heavily on deep learning applied to retinal fundus images, but the complex, irregular patterns of lesions in these images, which vary in shape and distribution, make it difficult to capture subtle changes. This study introduces RadFuse, a multi-representation deep learning framework that integrates non-linear RadEx-transformed sinogram images with traditional fundus images to enhance diabetic retinopathy detection and grading. Our RadEx transformation, an optimized non-linear extension of the Radon transform, generates sinogram representations to capture complex retinal lesion patterns. By leveraging both spatial and transformed domain information, RadFuse enriches the feature set available to deep learning models, improving the differentiation of severity levels. We conducted extensive experiments on two benchmark datasets, APTOS-2019 and DDR, using three convolutional neural networks (CNNs): ResNeXt-50, MobileNetV2, and VGG19. RadFuse showed significant improvements over fundus-image-only models across all three CNN architectures and outperformed state-of-the-art methods on both datasets. For severity grading across five stages, RadFuse achieved a quadratic weighted kappa of 93.24%, an accuracy of 87.07%, and an F1-score of 87.17%. In binary classification between healthy and diabetic retinopathy cases, the method reached an accuracy of 99.09%, precision of 98.58%, and recall of 99.6%, surpassing previously established models. These results demonstrate RadFuse's capacity to capture complex non-linear features, advancing diabetic retinopathy classification and promoting the integration of advanced mathematical transforms in medical image analysis.
Authors: Chengbo Zheng, Tim Miller, Alina Bialkowski, H Peter Soyer, Monika Janda
Abstract: High stakes decision-making often requires a continuous interplay between evolving evidence and shifting hypotheses, a dynamic that is not well supported by current AI decision support systems. In this paper, we introduce a mixed-initiative framework for AI assisted decision making that is grounded in the data-frame theory of sensemaking and the evaluative AI paradigm. Our approach enables both humans and AI to collaboratively construct, validate, and adapt hypotheses. We demonstrate our framework with an AI-assisted skin cancer diagnosis prototype that leverages a concept bottleneck model to facilitate interpretable interactions and dynamic updates to diagnostic hypotheses.
Authors: Chenxu Yang, Qingyi Si, Yongjie Duan, Zheliang Zhu, Chenyu Zhu, Zheng Lin, Li Cao, Weiping Wang
Abstract: Recent advances in large reasoning language models (LRLMs) rely on test-time scaling, which extends long chain-of-thought (CoT) generation to solve complex tasks. However, overthinking in long CoT not only slows down the efficiency of problem solving, but also risks accuracy loss due to the extremely detailed or redundant reasoning steps. We propose a simple yet effective method that allows LLMs to self-truncate CoT sequences by early exit during generation. Instead of relying on fixed heuristics, the proposed method monitors model behavior at potential reasoning transition points (e.g.,"Wait" tokens) and dynamically terminates the next reasoning chain's generation when the model exhibits high confidence in a trial answer. Our method requires no additional training and can be seamlessly integrated into existing o1-like reasoning LLMs. Experiments on multiple reasoning benchmarks MATH-500, AMC 2023, GPQA Diamond and AIME 2024 show that the proposed method is consistently effective on deepseek-series reasoning LLMs, reducing the length of CoT sequences by an average of 31% to 43% while improving accuracy by 1.7% to 5.7%.
Authors: Wenjing Xiao, Chenglong Shi, Miaojiang Chen, Zhiquan Liu, Min Chen, H. Herbert Song
Abstract: With the exponential growth of Internet of Things (IoT) devices, edge computing (EC) is gradually playing an important role in providing cost-effective services. However, existing approaches struggle to perform well in graph-structured scenarios where user data is correlated, such as traffic flow prediction and social relationship recommender systems. In particular, graph neural network (GNN)-based approaches lead to expensive server communication cost. To address this problem, we propose GraphEdge, an efficient GNN-based EC architecture. It considers the EC system of GNN tasks, where there are associations between users and it needs to take into account the task data of its neighbors when processing the tasks of a user. Specifically, the architecture first perceives the user topology and represents their data associations as a graph layout at each time step. Then the graph layout is optimized by calling our proposed hierarchical traversal graph cut algorithm (HiCut), which cuts the graph layout into multiple weakly associated subgraphs based on the aggregation characteristics of GNN, and the communication cost between different subgraphs during GNN inference is minimized. Finally, based on the optimized graph layout, our proposed deep reinforcement learning (DRL) based graph offloading algorithm (DRLGO) is executed to obtain the optimal offloading strategy for the tasks of users, the offloading strategy is subgraph-based, it tries to offload user tasks in a subgraph to the same edge server as possible while minimizing the task processing time and energy consumption of the EC system. Experimental results show the good effectiveness and dynamic adaptation of our proposed architecture and it also performs well even in dynamic scenarios.
Authors: Riley Pierson, Armin Moin
Abstract: Large open-source projects receive a large number of issues (known as bugs), including software defect (i.e., bug) reports and new feature requests from their user and developer communities at a fast rate. The often limited project resources do not allow them to deal with all issues. Instead, they have to prioritize them according to the project's priorities and the issues' severities. In this paper, we propose a novel approach to automated bug prioritization based on the natural language text of the bug reports that are stored in the open bug repositories of the issue-tracking systems. We conduct topic modeling using a variant of LDA called TopicMiner-MTM and text classification with the BERT large language model to achieve a higher performance level compared to the state-of-the-art. Experimental results using an existing reference dataset containing 85,156 bug reports of the Eclipse Platform project indicate that we outperform existing approaches in terms of Accuracy, Precision, Recall, and F1-measure of the bug report priority prediction.
Authors: Chang Zong, Bin Li, Shoujun Zhou, Jian Wan, Lei Zhang
Abstract: Locating specific segments within an instructional video is an efficient way to acquire guiding knowledge. Generally, the task of obtaining video segments for both verbal explanations and visual demonstrations is known as visual answer localization (VAL). However, users often need multiple interactions to obtain answers that align with their expectations when using the system. During these interactions, humans deepen their understanding of the video content by asking themselves questions, thereby accurately identifying the location. Therefore, we propose a new task, named In-VAL, to simulate the multiple interactions between humans and videos in the procedure of obtaining visual answers. The In-VAL task requires interactively addressing several semantic gap issues, including 1) the ambiguity of user intent in the input questions, 2) the incompleteness of language in video subtitles, and 3) the fragmentation of content in video segments. To address these issues, we propose Ask2Loc, a framework for resolving In-VAL by asking questions. It includes three key modules: 1) a chatting module to refine initial questions and uncover clear intentions, 2) a rewriting module to generate fluent language and create complete descriptions, and 3) a searching module to broaden local context and provide integrated content. We conduct extensive experiments on three reconstructed In-VAL datasets. Compared to traditional end-to-end and two-stage methods, our proposed Ask2Loc can improve performance by up to 14.91 (mIoU) on the In-VAL task. Our code and datasets can be accessed at https://github.com/changzong/Ask2Loc.
Authors: Alycia Carey, Xintao Wu
Abstract: Client-level fairness metrics for federated learning are used to ensure that all clients in a federation either: a) have similar final performance on their local data distributions (i.e., client parity), or b) obtain final performance on their local data distributions relative to their contribution to the federated learning process (i.e., contribution fairness). While a handful of works that propose either client-parity or contribution-based fairness metrics ground their definitions and decisions in social theories of equality -- such as distributive justice -- most works arbitrarily choose what notion of fairness to align with which makes it difficult for practitioners to choose which fairness metric aligns best with their fairness ethics. In this work, we propose UDJ-FL (Uncertainty-based Distributive Justice for Federated Learning), a flexible federated learning framework that can achieve multiple distributive justice-based client-level fairness metrics. Namely, by utilizing techniques inspired by fair resource allocation, in conjunction with performing aleatoric uncertainty-based client weighing, our UDJ-FL framework is able to achieve egalitarian, utilitarian, Rawls' difference principle, or desert-based client-level fairness. We empirically show the ability of UDJ-FL to achieve all four defined distributive justice-based client-level fairness metrics in addition to providing fairness equivalent to (or surpassing) other popular fair federated learning works. Further, we provide justification for why aleatoric uncertainty weighing is necessary to the construction of our UDJ-FL framework as well as derive theoretical guarantees for the generalization bounds of UDJ-FL. Our code is publicly available at https://github.com/alycia-noel/UDJ-FL.
Authors: Ling Cheng, Jiashu Pu, Ruicheng Liang, Qian Shao, Hezhe Qiao, Feida Zhu
Abstract: Semi-supervised community detection methods are widely used for identifying specific communities due to the label scarcity. Existing semi-supervised community detection methods typically involve two learning stages learning in both initial identification and subsequent adjustment, which often starts from an unreasonable community core candidate. Moreover, these methods encounter scalability issues because they depend on reinforcement learning and generative adversarial networks, leading to higher computational costs and restricting the selection of candidates. To address these limitations, we draw a parallel between crystallization kinetics and community detection to integrate the spontaneity of the annealing process into community detection. Specifically, we liken community detection to identifying a crystal subgrain (core) that expands into a complete grain (community) through a process similar to annealing. Based on this finding, we propose CLique ANNealing (CLANN), which applies kinetics concepts to community detection by integrating these principles into the optimization process to strengthen the consistency of the community core. Subsequently, a learning-free Transitive Annealer was employed to refine the first-stage candidates by merging neighboring cliques and repositioning the community core, enabling a spontaneous growth process that enhances scalability. Extensive experiments on \textbf{43} different network settings demonstrate that CLANN outperforms state-of-the-art methods across multiple real-world datasets, showcasing its exceptional efficacy and efficiency in community detection.
Authors: Meng Wang, Tian Lin, Qingshan Hou, Aidi Lin, Jingcheng Wang, Qingsheng Peng, Truong X. Nguyen, Danqi Fang, Ke Zou, Ting Xu, Cancan Xue, Ten Cheer Quek, Qinkai Yu, Minxin Liu, Hui Zhou, Zixuan Xiao, Guiqin He, Huiyu Liang, Tingkun Shi, Man Chen, Linna Liu, Yuanyuan Peng, Lianyu Wang, Qiuming Hu, Junhong Chen, Zhenhua Zhang, Cheng Chen, Yitian Zhao, Dianbo Liu, Jianhua Wu, Xinjian Chen, Changqing Zhang, Triet Thanh Nguyen, Yanda Meng, Yalin Zheng, Yih Chung Tham, Carol Y. Cheung, Huazhu Fu, Haoyu Chen, Ching-Yu Cheng
Abstract: Artificial intelligence (AI) shows remarkable potential in medical imaging diagnostics, but current models typically require retraining when deployed across different clinical centers, limiting their widespread adoption. We introduce GlobeReady, a clinician-friendly AI platform that enables ocular disease diagnosis without retraining/fine-tuning or technical expertise. GlobeReady achieves high accuracy across imaging modalities: 93.9-98.5% for an 11-category fundus photo dataset and 87.2-92.7% for a 15-category OCT dataset. Through training-free local feature augmentation, it addresses domain shifts across centers and populations, reaching an average accuracy of 88.9% across five centers in China, 86.3% in Vietnam, and 90.2% in the UK. The built-in confidence-quantifiable diagnostic approach further boosted accuracy to 94.9-99.4% (fundus) and 88.2-96.2% (OCT), while identifying out-of-distribution cases at 86.3% (49 CFP categories) and 90.6% (13 OCT categories). Clinicians from multiple countries rated GlobeReady highly (average 4.6 out of 5) for its usability and clinical relevance. These results demonstrate GlobeReady's robust, scalable diagnostic capability and potential to support ophthalmic care without technical barriers.
Authors: Saban Ozturk, Melih B. Yilmaz, Muti Kara, M. Talat Yavuz, Aykut Ko\c{c}, Tolga \c{C}ukur
Abstract: Diagnostic imaging relies on interpreting both images and radiology reports, but the growing data volumes place significant pressure on medical experts, yielding increased errors and workflow backlogs. Medical vision-language models (med-VLMs) have emerged as a powerful framework to efficiently process multimodal imaging data, particularly in chest X-ray (CXR) evaluations, albeit their performance hinges on how well image and text representations are aligned. Existing alignment methods, predominantly based on contrastive learning, prioritize separation between disease classes over segregation of fine-grained pathology attributes like location, size or severity, leading to suboptimal representations. Here, we propose MedTrim (Meta-entity-driven Triplet mining), a novel method that enhances image-text alignment through multimodal triplet learning synergistically guided by disease class as well as adjectival and directional pathology descriptors. Unlike common alignment methods that separate broad disease classes, MedTrim leverages structured meta-entity information to preserve subtle but clinically significant intra-class variations. For this purpose, we first introduce an ontology-based entity recognition module that extracts pathology-specific meta-entities from CXR reports, as annotations on pathology attributes are rare in public datasets. For refined sample selection in triplet mining, we then introduce a novel score function that captures an aggregate measure of inter-sample similarity based on disease classes and adjectival/directional descriptors. Lastly, we introduce a multimodal triplet alignment objective for explicit within- and cross-modal alignment between samples sharing detailed pathology characteristics. Our demonstrations indicate that MedTrim improves performance in downstream retrieval and classification tasks compared to state-of-the-art alignment methods.
Authors: Fanny Jourdan, Yannick Chevalier, C\'ecile Favre
Abstract: Large Language Models (LLMs) are increasingly leveraged for translation tasks but often fall short when translating inclusive language -- such as texts containing the singular 'they' pronoun or otherwise reflecting fair linguistic protocols. Because these challenges span both computational and societal domains, it is imperative to critically evaluate how well LLMs handle inclusive translation with a well-founded framework. This paper presents FairTranslate, a novel, fully human-annotated dataset designed to evaluate non-binary gender biases in machine translation systems from English to French. FairTranslate includes 2418 English-French sentence pairs related to occupations, annotated with rich metadata such as the stereotypical alignment of the occupation, grammatical gender indicator ambiguity, and the ground-truth gender label (male, female, or inclusive). We evaluate four leading LLMs (Gemma2-2B, Mistral-7B, Llama3.1-8B, Llama3.3-70B) on this dataset under different prompting procedures. Our results reveal substantial biases in gender representation across LLMs, highlighting persistent challenges in achieving equitable outcomes in machine translation. These findings underscore the need for focused strategies and interventions aimed at ensuring fair and inclusive language usage in LLM-based translation systems. We make the FairTranslate dataset publicly available on Hugging Face, and disclose the code for all experiments on GitHub.
Authors: Jerry Yao-Chieh Hu, Hude Liu, Hong-Yu Chen, Weimin Wu, Han Liu
Abstract: We prove that with linear transformations, both (i) two-layer self-attention and (ii) one-layer self-attention followed by a softmax function are universal approximators for continuous sequence-to-sequence functions on compact domains. Our main technique is a new interpolation-based method for analyzing attention's internal mechanism. This leads to our key insight: self-attention is able to approximate a generalized version of ReLU to arbitrary precision, and hence subsumes many known universal approximators. Building on these, we show that two-layer multi-head attention alone suffices as a sequence-to-sequence universal approximator. In contrast, prior works rely on feed-forward networks to establish universal approximation in Transformers. Furthermore, we extend our techniques to show that, (softmax-)attention-only layers are capable of approximating various statistical models in-context. We believe these techniques hold independent interest.
Authors: Sophie C. Pope, Andrew Barovic, Armin Moin
Abstract: This study explores a novel approach to predicting key bug-related outcomes, including the time to resolution, time to fix, and ultimate status of a bug, using data from the Bugzilla Eclipse Project. Specifically, we leverage features available before a bug is resolved to enhance predictive accuracy. Our methodology incorporates sentiment analysis to derive both an emotionality score and a sentiment classification (positive or negative). Additionally, we integrate the bug's priority level and its topic, extracted using a BERTopic model, as features for a Convolutional Neural Network (CNN) and a Multilayer Perceptron (MLP). Our findings indicate that the combination of BERTopic and sentiment analysis can improve certain model performance metrics. Furthermore, we observe that balancing model inputs enhances practical applicability, albeit at the cost of a significant reduction in accuracy in most cases. To address our primary objectives, predicting time-to-resolution, time-to-fix, and bug destiny, we employ both binary classification and exact time value predictions, allowing for a comparative evaluation of their predictive effectiveness. Results demonstrate that sentiment analysis serves as a valuable predictor of a bug's eventual outcome, particularly in determining whether it will be fixed. However, its utility is less pronounced when classifying bugs into more complex or unconventional outcome categories.
Authors: Shang Wang
Abstract: The demand for efficient natural language processing (NLP) systems has led to the development of lightweight language models. Previous work in this area has primarily focused on manual design or training-based neural architecture search (NAS) methods. Recently, zero-shot NAS methods have been proposed for evaluating language models without the need for training. However, prevailing approaches to zero-shot NAS often face challenges such as biased evaluation metrics and computational inefficiencies. In this paper, we introduce weight-weighted PCA (W-PCA), a novel zero-shot NAS method specifically tailored for lightweight language models. Our approach utilizes two evaluation proxies: the parameter count and the number of principal components with cumulative contribution exceeding $\eta$ in the feed-forward neural (FFN) layer. Additionally, by eliminating the need for gradient computations, we optimize the evaluation time, thus enhancing the efficiency of designing and evaluating lightweight language models. We conduct a comparative analysis on the GLUE and SQuAD datasets to evaluate our approach. The results demonstrate that our method significantly reduces training time compared to one-shot NAS methods and achieves higher scores in the testing phase compared to previous state-of-the-art training-based methods. Furthermore, we perform ranking evaluations on a dataset sampled from the FlexiBERT search space. Our approach exhibits superior ranking correlation and further reduces solving time compared to other zero-shot NAS methods that require gradient computation.
Authors: Sindhuja Madabushi, Ahmad Faraz Khan, Haider Ali, Jin-Hee Cho
Abstract: Vertical Federated Learning (VFL) enables organizations with disjoint feature spaces but shared user bases to collaboratively train models without sharing raw data. However, existing VFL systems face critical limitations: they often lack effective incentive mechanisms, struggle to balance privacy-utility tradeoffs, and fail to accommodate clients with heterogeneous resource capabilities. These challenges hinder meaningful participation, degrade model performance, and limit practical deployment. To address these issues, we propose OPUS-VFL, an Optimal Privacy-Utility tradeoff Strategy for VFL. OPUS-VFL introduces a novel, privacy-aware incentive mechanism that rewards clients based on a principled combination of model contribution, privacy preservation, and resource investment. It employs a lightweight leave-one-out (LOO) strategy to quantify feature importance per client, and integrates an adaptive differential privacy mechanism that enables clients to dynamically calibrate noise levels to optimize their individual utility. Our framework is designed to be scalable, budget-balanced, and robust to inference and poisoning attacks. Extensive experiments on benchmark datasets (MNIST, CIFAR-10, and CIFAR-100) demonstrate that OPUS-VFL significantly outperforms state-of-the-art VFL baselines in both efficiency and robustness. It reduces label inference attack success rates by up to 20%, increases feature inference reconstruction error (MSE) by over 30%, and achieves up to 25% higher incentives for clients that contribute meaningfully while respecting privacy and cost constraints. These results highlight the practicality and innovation of OPUS-VFL as a secure, fair, and performance-driven solution for real-world VFL.
Authors: Soham Bonnerjee (Kingsley), Zhen Wei (Kingsley), Yeon, Anna Asch, Sagnik Nandy, Promit Ghosal
Abstract: In-context learning (ICL)-the ability of transformer-based models to perform new tasks from examples provided at inference time-has emerged as a hallmark of modern language models. While recent works have investigated the mechanisms underlying ICL, its feasibility under formal privacy constraints remains largely unexplored. In this paper, we propose a differentially private pretraining algorithm for linear attention heads and present the first theoretical analysis of the privacy-accuracy trade-off for ICL in linear regression. Our results characterize the fundamental tension between optimization and privacy-induced noise, formally capturing behaviors observed in private training via iterative methods. Additionally, we show that our method is robust to adversarial perturbations of training prompts, unlike standard ridge regression. All theoretical findings are supported by extensive simulations across diverse settings.
Authors: Tom Zehle, Moritz Schlager, Timo Hei{\ss}, Matthias Feurer
Abstract: Large language models (LLMs) have revolutionized natural language processing by solving a wide range of tasks simply guided by a prompt. Yet their performance is highly sensitive to prompt formulation. While automated prompt optimization addresses this challenge by finding optimal prompts, current methods require a substantial number of LLM calls and input tokens, making prompt optimization expensive. We introduce CAPO (Cost-Aware Prompt Optimization), an algorithm that enhances prompt optimization efficiency by integrating AutoML techniques. CAPO is an evolutionary approach with LLMs as operators, incorporating racing to save evaluations and multi-objective optimization to balance performance with prompt length. It jointly optimizes instructions and few-shot examples while leveraging task descriptions for improved robustness. Our extensive experiments across diverse datasets and LLMs demonstrate that CAPO outperforms state-of-the-art discrete prompt optimization methods in 11/15 cases with improvements up to 21%p. Our algorithm achieves better performances already with smaller budgets, saves evaluations through racing, and decreases average prompt length via a length penalty, making it both cost-efficient and cost-aware. Even without few-shot examples, CAPO outperforms its competitors and generally remains robust to initial prompts. CAPO represents an important step toward making prompt optimization more powerful and accessible by improving cost-efficiency.
Authors: Soham Sane
Abstract: We introduce AlphaGrad, a memory-efficient, conditionally stateless optimizer addressing the memory overhead and hyperparameter complexity of adaptive methods like Adam. AlphaGrad enforces scale invariance via tensor-wise L2 gradient normalization followed by a smooth hyperbolic tangent transformation, $g' = \tanh(\alpha \cdot \tilde{g})$, controlled by a single steepness parameter $\alpha$. Our contributions include: (1) the AlphaGrad algorithm formulation; (2) a formal non-convex convergence analysis guaranteeing stationarity; (3) extensive empirical evaluation on diverse RL benchmarks (DQN, TD3, PPO). Compared to Adam, AlphaGrad demonstrates a highly context-dependent performance profile. While exhibiting instability in off-policy DQN, it provides enhanced training stability with competitive results in TD3 (requiring careful $\alpha$ tuning) and achieves substantially superior performance in on-policy PPO. These results underscore the critical importance of empirical $\alpha$ selection, revealing strong interactions between the optimizer's dynamics and the underlying RL algorithm. AlphaGrad presents a compelling alternative optimizer for memory-constrained scenarios and shows significant promise for on-policy learning regimes where its stability and efficiency advantages can be particularly impactful.
Authors: Dinithi Dissanayake, Suranga Nanayakkara
Abstract: Flow theory describes an optimal cognitive state where individuals experience deep focus and intrinsic motivation when a task's difficulty aligns with their skill level. In AI-augmented reasoning, interventions that disrupt the state of cognitive flow can hinder rather than enhance decision-making. This paper proposes a context-aware cognitive augmentation framework that adapts interventions based on three key contextual factors: type, timing, and scale. By leveraging multimodal behavioral cues (e.g., gaze behavior, typing hesitation, interaction speed), AI can dynamically adjust cognitive support to maintain or restore flow. We introduce the concept of cognitive flow, an extension of flow theory in AI-augmented reasoning, where interventions are personalized, adaptive, and minimally intrusive. By shifting from static interventions to context-aware augmentation, our approach ensures that AI systems support deep engagement in complex decision-making and reasoning without disrupting cognitive immersion.
Authors: Konstantin F. Pilz, James Sanders, Robi Rahman, Lennart Heim
Abstract: Frontier AI development relies on powerful AI supercomputers, yet analysis of these systems is limited. We create a dataset of 500 AI supercomputers from 2019 to 2025 and analyze key trends in performance, power needs, hardware cost, ownership, and global distribution. We find that the computational performance of AI supercomputers has doubled every nine months, while hardware acquisition cost and power needs both doubled every year. The leading system in March 2025, xAI's Colossus, used 200,000 AI chips, had a hardware cost of \$7B, and required 300 MW of power, as much as 250,000 households. As AI supercomputers evolved from tools for science to industrial machines, companies rapidly expanded their share of total AI supercomputer performance, while the share of governments and academia diminished. Globally, the United States accounts for about 75% of total performance in our dataset, with China in second place at 15%. If the observed trends continue, the leading AI supercomputer in 2030 will achieve $2\times10^{22}$ 16-bit FLOP/s, use two million AI chips, have a hardware cost of \$200 billion, and require 9 GW of power. Our analysis provides visibility into the AI supercomputer landscape, allowing policymakers to assess key AI trends like resource needs, ownership, and national competitiveness.
Authors: Ahmed R. Sadik, Siddhata Govind
Abstract: Determining the most effective Large Language Model for code smell detection presents a complex challenge. This study introduces a structured methodology and evaluation matrix to tackle this issue, leveraging a curated dataset of code samples consistently annotated with known smells. The dataset spans four prominent programming languages Java, Python, JavaScript, and C++; allowing for cross language comparison. We benchmark two state of the art LLMs, OpenAI GPT 4.0 and DeepSeek-V3, using precision, recall, and F1 score as evaluation metrics. Our analysis covers three levels of detail: overall performance, category level performance, and individual code smell type performance. Additionally, we explore cost effectiveness by comparing the token based detection approach of GPT 4.0 with the pattern-matching techniques employed by DeepSeek V3. The study also includes a cost analysis relative to traditional static analysis tools such as SonarQube. The findings offer valuable guidance for practitioners in selecting an efficient, cost effective solution for automated code smell detection
Authors: Yazan Otoum, Arghavan Asad, Amiya Nayak
Abstract: The rapid expansion of IoT ecosystems introduces severe challenges in scalability, security, and real-time decision-making. Traditional centralized architectures struggle with latency, privacy concerns, and excessive resource consumption, making them unsuitable for modern large-scale IoT deployments. This paper presents a novel Federated Learning-driven Large Language Model (FL-LLM) framework, designed to enhance IoT system intelligence while ensuring data privacy and computational efficiency. The framework integrates Generative IoT (GIoT) models with a Gradient Sensing Federated Strategy (GSFS), dynamically optimizing model updates based on real-time network conditions. By leveraging a hybrid edge-cloud processing architecture, our approach balances intelligence, scalability, and security in distributed IoT environments. Evaluations on the IoT-23 dataset demonstrate that our framework improves model accuracy, reduces response latency, and enhances energy efficiency, outperforming traditional FL techniques (i.e., FedAvg, FedOpt). These findings highlight the potential of integrating LLM-powered federated learning into large-scale IoT ecosystems, paving the way for more secure, scalable, and adaptive IoT management solutions.
Authors: Amund Tveit, Bj{\o}rn Remseth, Arve Skogvold
Abstract: This paper investigates the impact of different optimizers on the grokking phenomenon, where models exhibit delayed generalization. We conducted experiments across seven numerical tasks (primarily modular arithmetic) using a modern Transformer architecture. The experimental configuration systematically varied the optimizer (Muon vs. AdamW) and the softmax activation function (standard softmax, stablemax, and sparsemax) to assess their combined effect on learning dynamics. Our empirical evaluation reveals that the Muon optimizer, characterized by its use of spectral norm constraints and second-order information, significantly accelerates the onset of grokking compared to the widely used AdamW optimizer. Specifically, Muon reduced the mean grokking epoch from 153.09 to 102.89 across all configurations, a statistically significant difference (t = 5.0175, p = 6.33e-08). This suggests that the optimizer choice plays a crucial role in facilitating the transition from memorization to generalization.
Authors: Frank Li, Hari Trivedi, Bardia Khosravi, Theo Dapamede, Mohammadreza Chavoshi, Abdulhameed Dere, Rohan Satya Isaac, Aawez Mansuri, Janice Newsome, Saptarshi Purkayastha, Judy Gichoya
Abstract: Foundation models, trained on vast amounts of data using self-supervised techniques, have emerged as a promising frontier for advancing artificial intelligence (AI) applications in medicine. This study evaluates three different vision-language foundation models (RAD-DINO, CheXagent, and BiomedCLIP) on their ability to capture fine-grained imaging features for radiology tasks. The models were assessed across classification, segmentation, and regression tasks for pneumothorax and cardiomegaly on chest radiographs. Self-supervised RAD-DINO consistently excelled in segmentation tasks, while text-supervised CheXagent demonstrated superior classification performance. BiomedCLIP showed inconsistent performance across tasks. A custom segmentation model that integrates global and local features substantially improved performance for all foundation models, particularly for challenging pneumothorax segmentation. The findings highlight that pre-training methodology significantly influences model performance on specific downstream tasks. For fine-grained segmentation tasks, models trained without text supervision performed better, while text-supervised models offered advantages in classification and interpretability. These insights provide guidance for selecting foundation models based on specific clinical applications in radiology.
Authors: Zhifan Ye, Kejing Xia, Yonggan Fu, Xin Dong, Jihoon Hong, Xiangchi Yuan, Shizhe Diao, Jan Kautz, Pavlo Molchanov, Yingyan Celine Lin
Abstract: State space models (SSMs) have emerged as an efficient alternative to Transformer models for language modeling, offering linear computational complexity and constant memory usage as context length increases. However, despite their efficiency in handling long contexts, recent studies have shown that SSMs, such as Mamba models, generally underperform compared to Transformers in long-context understanding tasks. To address this significant shortfall and achieve both efficient and accurate long-context understanding, we propose LongMamba, a training-free technique that significantly enhances the long-context capabilities of Mamba models. LongMamba builds on our discovery that the hidden channels in Mamba can be categorized into local and global channels based on their receptive field lengths, with global channels primarily responsible for long-context capability. These global channels can become the key bottleneck as the input context lengthens. Specifically, when input lengths largely exceed the training sequence length, global channels exhibit limitations in adaptively extend their receptive fields, leading to Mamba's poor long-context performance. The key idea of LongMamba is to mitigate the hidden state memory decay in these global channels by preventing the accumulation of unimportant tokens in their memory. This is achieved by first identifying critical tokens in the global channels and then applying token filtering to accumulate only those critical tokens. Through extensive benchmarking across synthetic and real-world long-context scenarios, LongMamba sets a new standard for Mamba's long-context performance, significantly extending its operational range without requiring additional training. Our code is available at https://github.com/GATECH-EIC/LongMamba.
Authors: Sangeet Khemlani, Tyler Tran, Nathaniel Gyory, Anthony M. Harrison, Wallace E. Lawson, Ravenna Thielstrom, Hunter Thompson, Taaren Singh, J. Gregory Trafton
Abstract: Vision language models (VLMs) are designed to extract relevant visuospatial information from images. Some research suggests that VLMs can exhibit humanlike scene understanding, while other investigations reveal difficulties in their ability to process relational information. To achieve widespread applicability, VLMs must perform reliably, yielding comparable competence across a wide variety of related tasks. We sought to test how reliable these architectures are at engaging in trivial spatial cognition, e.g., recognizing whether one object is left of another in an uncluttered scene. We developed a benchmark dataset -- TableTest -- whose images depict 3D scenes of objects arranged on a table, and used it to evaluate state-of-the-art VLMs. Results show that performance could be degraded by minor variations of prompts that use logically equivalent descriptions. These analyses suggest limitations in how VLMs may reason about spatial relations in real-world applications. They also reveal novel opportunities for bolstering image caption corpora for more efficient training and testing.
Authors: Long Lian, Yifan Ding, Yunhao Ge, Sifei Liu, Hanzi Mao, Boyi Li, Marco Pavone, Ming-Yu Liu, Trevor Darrell, Adam Yala, Yin Cui
Abstract: Generating detailed and accurate descriptions for specific regions in images and videos remains a fundamental challenge for vision-language models. We introduce the Describe Anything Model (DAM), a model designed for detailed localized captioning (DLC). DAM preserves both local details and global context through two key innovations: a focal prompt, which ensures high-resolution encoding of targeted regions, and a localized vision backbone, which integrates precise localization with its broader context. To tackle the scarcity of high-quality DLC data, we propose a Semi-supervised learning (SSL)-based Data Pipeline (DLC-SDP). DLC-SDP starts with existing segmentation datasets and expands to unlabeled web images using SSL. We introduce DLC-Bench, a benchmark designed to evaluate DLC without relying on reference captions. DAM sets new state-of-the-art on 7 benchmarks spanning keyword-level, phrase-level, and detailed multi-sentence localized image and video captioning.
Authors: Thomas Schmied, J\"org Bornschein, Jordi Grau-Moya, Markus Wulfmeier, Razvan Pascanu
Abstract: The success of Large Language Models (LLMs) has sparked interest in various agentic applications. A key hypothesis is that LLMs, leveraging common sense and Chain-of-Thought (CoT) reasoning, can effectively explore and efficiently solve complex domains. However, LLM agents have been found to suffer from sub-optimal exploration and the knowing-doing gap, the inability to effectively act on knowledge present in the model. In this work, we systematically study why LLMs perform sub-optimally in decision-making scenarios. In particular, we closely examine three prevalent failure modes: greediness, frequency bias, and the knowing-doing gap. We propose mitigation of these shortcomings by fine-tuning via Reinforcement Learning (RL) on self-generated CoT rationales. Our experiments across multi-armed bandits, contextual bandits, and Tic-tac-toe, demonstrate that RL fine-tuning enhances the decision-making abilities of LLMs by increasing exploration and narrowing the knowing-doing gap. Finally, we study both classic exploration mechanisms, such as $\epsilon$-greedy, and LLM-specific approaches, such as self-correction and self-consistency, to enable more effective fine-tuning of LLMs for decision-making.
Authors: Pivithuru Thejan Amarasinghe, Su Nguyen, Yuan Sun, Damminda Alahakoon
Abstract: Business optimisation has been used extensively to determine optimal solutions for challenging business operations. Problem formulation is an important part of business optimisation as it influences both the validity of solutions and the efficiency of the optimisation process. While different optimisation modelling languages have been developed, problem formulation is still not a trivial task and usually requires optimisation expertise and problem-domain knowledge. Recently, Large Language Models (LLMs) have demonstrated outstanding performance across different language-related tasks. Since problem formulation can be viewed as a translation task, there is a potential to leverage LLMs to automate problem formulation. However, developing an LLM for problem formulation is challenging, due to limited training data, and the complexity of real-world optimisation problems. Several prompt engineering methods have been proposed in the literature to automate problem formulation with LLMs. While the initial results are encouraging, the accuracy of formulations generated by these methods can still be significantly improved. In this paper, we present an LLM-based framework for automating problem formulation in business optimization. Our approach introduces a method for fine-tuning cost-efficient LLMs specifically tailored to specialized business optimization challenges. The experiment results demonstrate that our framework can generate accurate formulations for conventional and real-world business optimisation problems in production scheduling. Extensive analyses show the effectiveness and the convergence of the proposed fine-tuning method. The proposed method also shows very competitive performance when compared with the state-of-the-art prompt engineering methods in the literature when tested on general linear programming problems.
Authors: Isha Chaudhary, Vedaant V. Jain, Gagandeep Singh
Abstract: Large Language Models (LLMs) are increasingly deployed in safety-critical systems where they provide answers based on in-context information derived from knowledge bases. As LLMs are increasingly envisioned as superhuman agents, their proficiency in knowledge comprehension-extracting relevant information and reasoning over it to answer questions, a key facet of human intelligence-becomes crucial. However, existing evaluations of LLMs on knowledge comprehension are typically conducted on small test sets, but these datasets represent only a tiny fraction of the vast number of possible queries. Simple empirical evaluations on these limited test sets raises concerns about the reliability and generalizability of the results. In this work, we introduce the first specification and certification framework for knowledge comprehension in LLMs, providing formal probabilistic guarantees for reliability. Instead of a fixed dataset, we design novel specifications that mathematically represent prohibitively large probability distributions of knowledge comprehension prompts with natural noise, using knowledge graphs. From these specifications, we generate quantitative certificates that offer high-confidence, tight bounds on the probability that a given LLM correctly answers any question drawn from the specification distribution. We apply our framework to certify SOTA LLMs in two domains: precision medicine and general question-answering. Our results reveal previously unrecognized vulnerabilities in SOTA LLMs due to natural noise in the prompts. Additionally, we establish performance hierarchies with formal guarantees among the SOTA LLMs, particularly in the context of precision medicine question-answering.
Authors: Isha Chaudhary, Qian Hu, Manoj Kumar, Morteza Ziyadi, Rahul Gupta, Gagandeep Singh
Abstract: Large Language Models (LLMs) can produce biased responses that can cause representational harms. However, conventional studies are insufficient to thoroughly evaluate biases across LLM responses for different demographic groups (a.k.a. counterfactual bias), as they do not scale to large number of inputs and do not provide guarantees. Therefore, we propose the first framework, LLMCert-B that certifies LLMs for counterfactual bias on distributions of prompts. A certificate consists of high-confidence bounds on the probability of unbiased LLM responses for any set of counterfactual prompts - prompts differing by demographic groups, sampled from a distribution. We illustrate counterfactual bias certification for distributions of counterfactual prompts created by applying prefixes sampled from prefix distributions, to a given set of prompts. We consider prefix distributions consisting random token sequences, mixtures of manual jailbreaks, and perturbations of jailbreaks in LLM's embedding space. We generate non-trivial certificates for SOTA LLMs, exposing their vulnerabilities over distributions of prompts generated from computationally inexpensive prefix distributions.
Authors: Zhou Xiaochen, Liang Xingzhou, Zou Hui, Lu Yi, Qu Jingjing
Abstract: In this paper, we propose a method to automatically classify AI-related documents from large-scale literature databases, leading to the creation of an AI-related literature dataset, named DeepDiveAI. The dataset construction approach integrates expert knowledge with the capabilities of advanced models, structured across two global stages. In the first stage, expert-curated classification datasets are used to train an LSTM model, which classifies coarse AI related records from large-scale datasets. In the second stage, we use Qwen2.5 Plus to annotate a random 10% of the coarse AI-related records, which are then used to train a BERT binary classifier. This step further refines the coarse AI related record set to obtain the final DeepDiveAI dataset. Evaluation results demonstrate that the entire workflow can efficiently and accurately identify AI-related literature from large-scale datasets.
Authors: Marharyta Domnich, Julius V\"alja, Rasmus Moorits Veski, Giacomo Magnifico, Kadi Tulver, Eduard Barbu, Raul Vicente
Abstract: As machine learning models evolve, maintaining transparency demands more human-centric explainable AI techniques. Counterfactual explanations, with roots in human reasoning, identify the minimal input changes needed to obtain a given output and, hence, are crucial for supporting decision-making. Despite their importance, the evaluation of these explanations often lacks grounding in user studies and remains fragmented, with existing metrics not fully capturing human perspectives. To address this challenge, we developed a diverse set of 30 counterfactual scenarios and collected ratings across 8 evaluation metrics from 206 respondents. Subsequently, we fine-tuned different Large Language Models (LLMs) to predict average or individual human judgment across these metrics. Our methodology allowed LLMs to achieve an accuracy of up to 63% in zero-shot evaluations and 85% (over a 3-classes prediction) with fine-tuning across all metrics. The fine-tuned models predicting human ratings offer better comparability and scalability in evaluating different counterfactual explanation frameworks.
Authors: Fabrizio Davide, Pietro Torre, Leonardo Ercolani, Andrea Gaggioli
Abstract: We tasked 16 state-of-the-art large language models (LLMs) with estimating the likelihood of Artificial General Intelligence (AGI) emerging by 2030. To assess the quality of these forecasts, we implemented an automated peer review process (LLM-PR). The LLMs' estimates varied widely, ranging from 3% (Reka- Core) to 47.6% (GPT-4o), with a median of 12.5%. These estimates closely align with a recent expert survey that projected a 10% likelihood of AGI by 2027, underscoring the relevance of LLMs in forecasting complex, speculative scenarios. The LLM-PR process demonstrated strong reliability, evidenced by a high Intraclass Correlation Coefficient (ICC = 0.79), reflecting notable consistency in scoring across the models. Among the models, Pplx-70b-online emerged as the top performer, while Gemini-1.5-pro-api ranked the lowest. A cross-comparison with external benchmarks, such as LMSYS Chatbot Arena, revealed that LLM rankings remained consistent across different evaluation methods, suggesting that existing benchmarks may not encapsulate some of the skills relevant for AGI prediction. We further explored the use of weighting schemes based on external benchmarks, optimizing the alignment of LLMs' predictions with human expert forecasts. This analysis led to the development of a new, 'AGI benchmark' designed to highlight performance differences in AGI-related tasks. Our findings offer insights into LLMs' capabilities in speculative, interdisciplinary forecasting tasks and emphasize the growing need for innovative evaluation frameworks for assessing AI performance in complex, uncertain real-world scenarios.
Authors: Matthew Stephenson, Matthew Sidji, Beno\^it Ronval
Abstract: In this paper, we propose the use of the popular word-based board game Codenames as a suitable benchmark for evaluating the reasoning capabilities of Large Language Models (LLMs). Codenames presents a highly interesting challenge for achieving successful AI performance, requiring both a sophisticated understanding of language, theory of mind, and epistemic reasoning capabilities. Prior attempts to develop agents for Codenames have largely relied on word embedding techniques, which have a limited vocabulary range and perform poorly when paired with differing approaches. LLMs have demonstrated enhanced reasoning and comprehension capabilities for language-based tasks, but can still suffer in lateral thinking challenges. We evaluate the capabilities of several state-of-the-art LLMs, including GPT-4o, Gemini 1.5, Claude 3.5 Sonnet, and Llama 3.1, across a variety of board setups. Our results indicate that while certain LLMs perform better than others overall, different models exhibit varying emergent behaviours during gameplay and excel at specific roles. We also evaluate the performance of different combinations of LLMs when playing cooperatively together, demonstrating that LLM agents are more generalisable to a wider range of teammates than prior techniques.
Authors: Timoth\'ee Anne, Noah Syrkis, Meriem Elhosni, Florian Turati, Franck Legendre, Alain Jaquier, Sebastian Risi
Abstract: Large Language Models (LLMs) have demonstrated remarkable performance across various tasks. Their potential to facilitate human coordination with many agents is a promising but largely under-explored area. Such capabilities would be helpful in disaster response, urban planning, and real-time strategy scenarios. In this work, we introduce (1) a real-time strategy game benchmark designed to evaluate these abilities and (2) a novel framework we term HIVE. HIVE empowers a single human to coordinate swarms of up to 2,000 agents through a natural language dialog with an LLM. We present promising results on this multi-agent benchmark, with our hybrid approach solving tasks such as coordinating agent movements, exploiting unit weaknesses, leveraging human annotations, and understanding terrain and strategic points. Our findings also highlight critical limitations of current models, including difficulties in processing spatial visual information and challenges in formulating long-term strategic plans. This work sheds light on the potential and limitations of LLMs in human-swarm coordination, paving the way for future research in this area. The HIVE project page, hive.syrkis.com, includes videos of the system in action.
Authors: Zhixuan Pan, Shaowen Wang, Jian Li
Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities across numerous tasks, yet principled explanations for their underlying mechanisms and several phenomena, such as scaling laws, hallucinations, and related behaviors, remain elusive. In this work, we revisit the classical relationship between compression and prediction, grounded in Kolmogorov complexity and Shannon information theory, to provide deeper insights into LLM behaviors. By leveraging the Kolmogorov Structure Function and interpreting LLM compression as a two-part coding process, we offer a detailed view of how LLMs acquire and store information across increasing model and data scales -- from pervasive syntactic patterns to progressively rarer knowledge elements. Motivated by this theoretical perspective and natural assumptions inspired by Heap's and Zipf's laws, we introduce a simplified yet representative hierarchical data-generation framework called the Syntax-Knowledge model. Under the Bayesian setting, we show that prediction and compression within this model naturally lead to diverse learning and scaling behaviors of LLMs. In particular, our theoretical analysis offers intuitive and principled explanations for both data and model scaling laws, the dynamics of knowledge acquisition during training and fine-tuning, factual knowledge hallucinations in LLMs. The experimental results validate our theoretical predictions.
Authors: Christopher Zhang Cui, Xingdi Yuan, Ziang Xiao, Prithviraj Ammanabrolu, Marc-Alexandre C\^ot\'e
Abstract: Reasoning is an essential skill to enable Large Language Models (LLMs) to interact with the world. As tasks become more complex, they demand increasingly sophisticated and diverse reasoning capabilities for sequential decision-making, requiring structured reasoning over the context history to determine the next best action. We introduce TALES, a diverse collection of synthetic and human-written text-adventure games designed to challenge and evaluate diverse reasoning capabilities. We present results over a range of LLMs, open- and closed-weights, performing a qualitative analysis on the top performing models. Despite an impressive showing on synthetic games, even the top LLM-driven agents fail to achieve 15% on games designed for human enjoyment. Code and visualization of the experiments can be found at https://microsoft.github.io/tales.
Authors: Alessio Buscemi, Daniele Proverbio, Alessandro Di Stefano, The Anh Han, German Castignani, Pietro Li\`o
Abstract: Letting AI agents interact in multi-agent applications adds a layer of complexity to the interpretability and prediction of AI outcomes, with profound implications for their trustworthy adoption in research and society. Game theory offers powerful models to capture and interpret strategic interaction among agents, but requires the support of reproducible, standardized and user-friendly IT frameworks to enable comparison and interpretation of results. To this end, we present FAIRGAME, a Framework for AI Agents Bias Recognition using Game Theory. We describe its implementation and usage, and we employ it to uncover biased outcomes in popular games among AI agents, depending on the employed Large Language Model (LLM) and used language, as well as on the personality trait or strategic knowledge of the agents. Overall, FAIRGAME allows users to reliably and easily simulate their desired games and scenarios and compare the results across simulation campaigns and with game-theoretic predictions, enabling the systematic discovery of biases, the anticipation of emerging behavior out of strategic interplays, and empowering further research into strategic decision-making using LLM agents.
Authors: Yi Sun, Han Wang, Jiaqiang Li, Jiacheng Liu, Xiangyu Li, Hao Wen, Huiwen Zheng, Yan Liang, Yuanchun Li, Yunxin Liu
Abstract: Recent work has demonstrated the remarkable potential of Large Language Models (LLMs) in test-time scaling. By making the models think before answering, they are able to achieve much higher accuracy with extra inference computation. However, in many real-world scenarios, models are used under time constraints, where an answer should be given to the user within a certain output length. It is unclear whether and how the reasoning abilities of LLMs remain effective under such constraints. We take a first look at this problem by conducting an in-depth empirical study. Specifically, we test more than 25 LLMs on common reasoning datasets under a wide range of output length budgets, and we analyze the correlation between the inference accuracy and various properties including model type, model size, prompt style, etc. We also consider the mappings between the token budgets and the actual on-device latency budgets. The results have demonstrated several interesting findings regarding the budget-aware LLM reasoning that differ from the unconstrained situation, e.g. the optimal choices of model sizes and prompts change under different budgets. These findings offer practical guidance for users to deploy LLMs under real-world latency constraints.
Authors: Shin-nosuke Ishikawa, Atsushi Yoshino
Abstract: The human-level performance of Large Language Models (LLMs) across various tasks has raised expectations for the potential of Artificial Intelligence (AI) to possess emotions someday. To explore the capability of current LLMs to express emotions in their outputs, we conducted an experiment using several LLMs (OpenAI GPT, Google Gemini, Meta Llama3, and Cohere Command R+) to role-play as agents answering questions with specified emotional states. We defined the emotional states using Russell's Circumplex model, a well-established framework that characterizes emotions along the sleepy-activated (arousal) and pleasure-displeasure (valence) axes. We chose this model for its simplicity, utilizing two continuous parameters, which allows for better controllability in applications involving continuous changes in emotional states. The responses generated were evaluated using a sentiment analysis model, independent of the LLMs, trained on the GoEmotions dataset. The evaluation showed that the emotional states of the generated answers were consistent with the specifications, demonstrating the LLMs' capability for emotional expression. This indicates the potential for LLM-based AI agents to simulate emotions, opening up a wide range of applications for emotion-based interactions, such as advisors or consultants who can provide advice or opinions with a personal touch.
Authors: Shilin Zhang, Zican Hu, Wenhao Wu, Xinyi Xie, Jianxiang Tang, Chunlin Chen, Daoyi Dong, Yu Cheng, Zhenhong Sun, Zhi Wang
Abstract: RL systems usually tackle generalization by inferring task beliefs from high-quality samples or warmup explorations. The restricted form limits their generality and usability since these supervision signals are expensive and even infeasible to acquire in advance for unseen tasks. Learning directly from the raw text about decision tasks is a promising alternative to leverage a much broader source of supervision. In the paper, we propose Text-to-Decision Agent (T2DA), a simple and scalable framework that supervises generalist policy learning with natural language. We first introduce a generalized world model to encode multi-task decision data into a dynamics-aware embedding space. Then, inspired by CLIP, we predict which textual description goes with which decision embedding, effectively bridging their semantic gap via contrastive language-decision pre-training and aligning the text embeddings to comprehend the environment dynamics. After training the text-conditioned generalist policy, the agent can directly realize zero-shot text-to-decision generation in response to language instructions. Comprehensive experiments on MuJoCo and Meta-World benchmarks show that T2DA facilitates high-capacity zero-shot generalization and outperforms various types of baselines.
Authors: Yizhu Jiao, Xuchao Zhang, Zhaoyang Wang, Yubo Ma, Zhun Deng, Rujia Wang, Chetan Bansal, Saravan Rajmohan, Jiawei Han, Huaxiu Yao
Abstract: Current Large Language Models (LLMs) excel in general reasoning yet struggle with specialized tasks requiring proprietary or domain-specific knowledge. Fine-tuning large models for every niche application is often infeasible due to black-box constraints and high computational overhead. To address this, we propose a collaborative framework that pairs a specialized weak model with a general strong model. The weak model, tailored to specific domains, produces initial drafts and background information, while the strong model leverages its advanced reasoning to refine these drafts, extending LLMs' capabilities to critical yet specialized tasks. To optimize this collaboration, we introduce a collaborative feedback to fine-tunes the weak model, which quantifies the influence of the weak model's contributions in the collaboration procedure and establishes preference pairs to guide preference tuning of the weak model. We validate our framework through experiments on three domains. We find that the collaboration significantly outperforms each model alone by leveraging complementary strengths. Moreover, aligning the weak model with the collaborative preference further enhances overall performance.
Authors: Monan Zhou, Xiaobing Li, Feng Yu, Wei Li
Abstract: The EMelodyGen system focuses on emotional melody generation in ABC notation controlled by the musical feature template. Owing to the scarcity of well-structured and emotionally labeled sheet music, we designed a template for controlling emotional melody generation by statistical correlations between musical features and emotion labels derived from small-scale emotional symbolic music datasets and music psychology conclusions. We then automatically annotated a large, well-structured sheet music collection with rough emotional labels by the template, converted them into ABC notation, and reduced label imbalance by data augmentation, resulting in a dataset named Rough4Q. Our system backbone pre-trained on Rough4Q can achieve up to 99% music21 parsing rate and melodies generated by our template can lead to a 91% alignment on emotional expressions in blind listening tests. Ablation studies further validated the effectiveness of the feature controls in the template. Available code and demos are at https://github.com/monetjoe/EMelodyGen.
Authors: Jun Wang, Hosein Hasanbeig, Kaiyuan Tan, Zihe Sun, Yiannis Kantaros
Abstract: This paper addresses the problem of designing control policies for agents with unknown stochastic dynamics and control objectives specified using Linear Temporal Logic (LTL). Recent Deep Reinforcement Learning (DRL) algorithms have aimed to compute policies that maximize the satisfaction probability of LTL formulas, but they often suffer from slow learning performance. To address this, we introduce a novel Deep Q-learning algorithm that significantly improves learning speed. The enhanced sample efficiency stems from a mission-driven exploration strategy that prioritizes exploration towards directions likely to contribute to mission success. Identifying these directions relies on an automaton representation of the LTL task as well as a learned neural network that partially models the agent-environment interaction. We provide comparative experiments demonstrating the efficiency of our algorithm on robot navigation tasks in unseen environments.
Authors: Asad Aali, Giannis Daras, Brett Levac, Sidharth Kumar, Alexandros G. Dimakis, Jonathan I. Tamir
Abstract: We provide a framework for solving inverse problems with diffusion models learned from linearly corrupted data. Firstly, we extend the Ambient Diffusion framework to enable training directly from measurements corrupted in the Fourier domain. Subsequently, we train diffusion models for MRI with access only to Fourier subsampled multi-coil measurements at acceleration factors R= 2,4,6,8. Secondly, we propose Ambient Diffusion Posterior Sampling (A-DPS), a reconstruction algorithm that leverages generative models pre-trained on one type of corruption (e.g. image inpainting) to perform posterior sampling on measurements from a different forward process (e.g. image blurring). For MRI reconstruction in high acceleration regimes, we observe that A-DPS models trained on subsampled data are better suited to solving inverse problems than models trained on fully sampled data. We also test the efficacy of A-DPS on natural image datasets (CelebA, FFHQ, and AFHQ) and show that A-DPS can sometimes outperform models trained on clean data for several image restoration tasks in both speed and performance.
Authors: Yui Lo, Yuqian Chen, Dongnan Liu, Wan Liu, Leo Zekelman, Fan Zhang, Yogesh Rathi, Nikos Makris, Alexandra J. Golby, Weidong Cai, Lauren J. O'Donnell
Abstract: Shape plays an important role in computer graphics, offering informative features to convey an object's morphology and functionality. Shape analysis in brain imaging can help interpret structural and functionality correlations of the human brain. In this work, we investigate the shape of the brain's 3D white matter connections and its potential predictive relationship to human cognitive function. We reconstruct brain connections as sequences of 3D points using diffusion magnetic resonance imaging (dMRI) tractography. To describe each connection, we extract 12 shape descriptors in addition to traditional dMRI connectivity and tissue microstructure features. We introduce a novel framework, Shape--fused Fiber Cluster Transformer (SFFormer), that leverages a multi-head cross-attention feature fusion module to predict subject-specific language performance based on dMRI tractography. We assess the performance of the method on a large dataset including 1065 healthy young adults. The results demonstrate that both the transformer-based SFFormer model and its inter/intra feature fusion with shape, microstructure, and connectivity are informative, and together, they improve the prediction of subject-specific language performance scores. Overall, our results indicate that the shape of the brain's connections is predictive of human language function.
Authors: Runze Lin, Junghui Chen, Lei Xie, Hongye Su
Abstract: In the context of Industry 4.0 and smart manufacturing, the field of process industry optimization and control is also undergoing a digital transformation. With the rise of Deep Reinforcement Learning (DRL), its application in process control has attracted widespread attention. However, the extremely low sample efficiency and the safety concerns caused by exploration in DRL hinder its practical implementation in industrial settings. Transfer learning offers an effective solution for DRL, enhancing its generalization and adaptability in multi-mode control scenarios. This paper provides insights into the use of DRL for process control from the perspective of transfer learning. We analyze the challenges of applying DRL in the process industry and the necessity of introducing transfer learning. Furthermore, recommendations and prospects are provided for future research directions on how transfer learning can be integrated with DRL to enhance process control. This paper aims to offer a set of promising, user-friendly, easy-to-implement, and scalable approaches to artificial intelligence-facilitated industrial control for scholars and engineers in the process industry.
Authors: Shehu AbdusSalam, Steve Abel, Miguel Crispim Romao
Abstract: We propose symbolic regression as a powerful tool for studying Beyond the Standard Model physics. As a benchmark model, we consider the so-called Constrained Minimal Supersymmetric Standard Model, which has a four-dimensional parameter space defined at the GUT scale. We provide a set of analytical expressions that reproduce three low-energy observables of interest in terms of the parameters of the theory: the Higgs mass, the contribution to the anomalous magnetic moment of the muon, and the cold dark matter relic density. To demonstrate the power of the approach, we employ the symbolic expressions in a global fits analysis to derive the posterior probability densities of the parameters, which are obtained extremely rapidly in comparison with conventional methods.
Authors: Quang H. Nguyen, Thinh Dao, Duy C. Hoang, Juliette Decugis, Saurav Manchanda, Nitesh V. Chawla, Khoa D. Doan
Abstract: The rapid progress in machine learning (ML) has brought forth many large language models (LLMs) that excel in various tasks and areas. These LLMs come with different abilities and costs in terms of computation or pricing. Since the demand for each query can vary, e.g., because of the queried domain or its complexity, defaulting to one LLM in an application is not usually the best choice, whether it is the biggest, priciest, or even the one with the best average test performance. Consequently, picking the right LLM that is both accurate and cost-effective for an application is necessary yet remains a challenge. In this paper, we introduce MetaLLM, a framework that dynamically and intelligently routes each query to the optimal LLM (among several available LLMs) for classification and multi-choice question-answering tasks, achieving significantly improved accuracy and cost-effectiveness. By framing the selection problem as a multi-armed bandit, MetaLLM balances prediction accuracy and cost efficiency under uncertainty. Our experiments, conducted on popular LLM platforms such as OpenAI and Together AI, as well as open-source LLM, showcase MetaLLM's efficacy in real-world scenarios, laying the groundwork for future extensions.
Authors: Yucong Zhang, Xin Zou, Jinshan Yang, Wenjun Chen, Juan Liu, Faya Liang, Ming Li
Abstract: This paper presents the Multimodal Laryngoscopic Video Analyzing System (MLVAS), a novel system that leverages both audio and video data to automatically extract key video segments and metrics from raw laryngeal videostroboscopic videos for assisted clinical assessment. The system integrates video-based glottis detection with an audio keyword spotting method to analyze both video and audio data, identifying patient vocalizations and refining video highlights to ensure optimal inspection of vocal fold movements. Beyond key video segment extraction from the raw laryngeal videos, MLVAS is able to generate effective audio and visual features for Vocal Fold Paralysis (VFP) detection. Pre-trained audio encoders are utilized to encode the patient voice to get the audio features. Visual features are generated by measuring the angle deviation of both the left and right vocal folds to the estimated glottal midline on the segmented glottis masks. To get better masks, we introduce a diffusion-based refinement that follows traditional U-Net segmentation to reduce false positives. We conducted several ablation studies to demonstrate the effectiveness of each module and modalities in the proposed MLVAS. The experimental results on a public segmentation dataset show the effectiveness of our proposed segmentation module. In addition, unilateral VFP classification results on a real-world clinic dataset demonstrate MLVAS's ability of providing reliable and objective metrics as well as visualization for assisted clinical diagnosis.
Authors: Rongfeng Lu, Hangyu Chen, Zunjie Zhu, Yuhang Qin, Ming Lu, Le Zhang, Chenggang Yan, Anke Xue
Abstract: Thermography is especially valuable for the military and other users of surveillance cameras. Some recent methods based on Neural Radiance Fields (NeRF) are proposed to reconstruct the thermal scenes in 3D from a set of thermal and RGB images. However, unlike NeRF, 3D Gaussian splatting (3DGS) prevails due to its rapid training and real-time rendering. In this work, we propose ThermalGaussian, the first thermal 3DGS approach capable of rendering high-quality images in RGB and thermal modalities. We first calibrate the RGB camera and the thermal camera to ensure that both modalities are accurately aligned. Subsequently, we use the registered images to learn the multimodal 3D Gaussians. To prevent the overfitting of any single modality, we introduce several multimodal regularization constraints. We also develop smoothing constraints tailored to the physical characteristics of the thermal modality. Besides, we contribute a real-world dataset named RGBT-Scenes, captured by a hand-hold thermal-infrared camera, facilitating future research on thermal scene reconstruction. We conduct comprehensive experiments to show that ThermalGaussian achieves photorealistic rendering of thermal images and improves the rendering quality of RGB images. With the proposed multimodal regularization constraints, we also reduced the model's storage cost by 90%. Our project page is at https://thermalgaussian.github.io/.
Authors: Junfeng Fang, Houcheng Jiang, Kun Wang, Yunshan Ma, Shi Jie, Xiang Wang, Xiangnan He, Tat-seng Chua
Abstract: Large language models (LLMs) often exhibit hallucinations due to incorrect or outdated knowledge. Hence, model editing methods have emerged to enable targeted knowledge updates. To achieve this, a prevailing paradigm is the locating-then-editing approach, which first locates influential parameters and then edits them by introducing a perturbation. While effective, current studies have demonstrated that this perturbation inevitably disrupt the originally preserved knowledge within LLMs, especially in sequential editing scenarios. To address this, we introduce AlphaEdit, a novel solution that projects perturbation onto the null space of the preserved knowledge before applying it to the parameters. We theoretically prove that this projection ensures the output of post-edited LLMs remains unchanged when queried about the preserved knowledge, thereby mitigating the issue of disruption. Extensive experiments on various LLMs, including LLaMA3, GPT2-XL, and GPT-J, show that AlphaEdit boosts the performance of most locating-then-editing methods by an average of 36.7% with a single line of additional code for projection solely. Our code is available at: https://github.com/jianghoucheng/AlphaEdit.
Authors: Attila Lovas
Abstract: Nonlinear time series models with exogenous regressors are essential in econometrics, queuing theory, and machine learning, though their statistical analysis remains incomplete. Key results, such as the law of large numbers and the functional central limit theorem, are known for weakly dependent variables. We demonstrate the transfer of mixing properties from the exogenous regressor to the response via coupling arguments. Additionally, we study Markov chains in random environments with drift and minorization conditions, even under non-stationary environments with favorable mixing properties, and apply this framework to single-server queuing models.
Authors: Xiaogeng Liu, Peiran Li, Edward Suh, Yevgeniy Vorobeychik, Zhuoqing Mao, Somesh Jha, Patrick McDaniel, Huan Sun, Bo Li, Chaowei Xiao
Abstract: In this paper, we propose AutoDAN-Turbo, a black-box jailbreak method that can automatically discover as many jailbreak strategies as possible from scratch, without any human intervention or predefined scopes (e.g., specified candidate strategies), and use them for red-teaming. As a result, AutoDAN-Turbo can significantly outperform baseline methods, achieving a 74.3% higher average attack success rate on public benchmarks. Notably, AutoDAN-Turbo achieves an 88.5 attack success rate on GPT-4-1106-turbo. In addition, AutoDAN-Turbo is a unified framework that can incorporate existing human-designed jailbreak strategies in a plug-and-play manner. By integrating human-designed strategies, AutoDAN-Turbo can even achieve a higher attack success rate of 93.4 on GPT-4-1106-turbo.
Authors: Sam Gunn, Xuandong Zhao, Dawn Song
Abstract: We present the first undetectable watermarking scheme for generative image models. Undetectability ensures that no efficient adversary can distinguish between watermarked and un-watermarked images, even after making many adaptive queries. In particular, an undetectable watermark does not degrade image quality under any efficiently computable metric. Our scheme works by selecting the initial latents of a diffusion model using a pseudorandom error-correcting code (Christ and Gunn, 2024), a strategy which guarantees undetectability and robustness. We experimentally demonstrate that our watermarks are quality-preserving and robust using Stable Diffusion 2.1. Our experiments verify that, in contrast to every prior scheme we tested, our watermark does not degrade image quality. Our experiments also demonstrate robustness: existing watermark removal attacks fail to remove our watermark from images without significantly degrading the quality of the images. Finally, we find that we can robustly encode 512 bits in our watermark, and up to 2500 bits when the images are not subjected to watermark removal attacks. Our code is available at https://github.com/XuandongZhao/PRC-Watermark.
Authors: Hui Yuan, Yifan Zeng, Yue Wu, Huazheng Wang, Mengdi Wang, Liu Leqi
Abstract: Reinforcement Learning from Human Feedback (RLHF) has become the predominant approach for language model (LM) alignment. At its core, RLHF uses a margin-based loss for preference optimization, specifying ideal LM behavior only by the difference between preferred and dispreferred responses. In this paper, we identify a common pitfall of margin-based methods -- the under-specification of ideal LM behavior on preferred and dispreferred responses individually, which leads to two unintended consequences as the margin increases: (1) The probability of dispreferred (e.g., unsafe) responses may increase, resulting in potential safety alignment failures. (2) The probability of preferred responses may decrease, even when those responses are ideal. We demystify the reasons behind these problematic behaviors: margin-based losses couple the change in the preferred probability to the gradient of the dispreferred one, and vice versa, often preventing the preferred probability from increasing while the dispreferred one decreases, and thus causing a synchronized increase or decrease in both probabilities. We term this effect, inherent in margin-based objectives, gradient entanglement. Formally, we derive conditions for general margin-based alignment objectives under which gradient entanglement becomes concerning: the inner product of the gradients of preferred and dispreferred log-probabilities is large relative to the individual gradient norms. We theoretically investigate why such inner products can be large when aligning language models and empirically validate our findings. Empirical implications of our framework extend to explaining important differences in the training dynamics of various preference optimization algorithms, and suggesting potential algorithm designs to mitigate the under-specification issue of margin-based methods and thereby improving language model alignment.
Authors: Yanjun Chen, Xinming Zhang, Xianghui Wang, Zhiqiang Xu, Xiaoyu Shen, Wei Zhang
Abstract: Soft Actor-Critic algorithm is widely recognized for its robust performance across a range of deep reinforcement learning tasks, where it leverages the tanh transformation to constrain actions within bounded limits. However, this transformation induces a distribution shift, distorting the original Gaussian action distribution and potentially leading the policy to select suboptimal actions, particularly in high-dimensional action spaces. In this paper, we conduct a comprehensive theoretical and empirical analysis of this distribution shift, deriving the precise probability density function (PDF) for actions following the tanh transformation to clarify the misalignment introduced between the transformed distribution's mode and the intended action output. We substantiate these theoretical insights through extensive experiments on high-dimensional tasks within the HumanoidBench benchmark. Our findings indicate that accounting for this distribution shift substantially enhances SAC's performance, resulting in notable improvements in cumulative rewards, sample efficiency, and reliability across tasks. These results underscore a critical consideration for SAC and similar algorithms: addressing transformation-induced distribution shifts is essential to optimizing policy effectiveness in high-dimensional deep reinforcement learning environments, thereby expanding the robustness and applicability of SAC in complex control tasks.
Authors: Yifu Sun, Xulong Zhang, Monan Zhou, Wei Li
Abstract: Music emotion recognition (MER) aims to identify the emotions conveyed in a given musical piece. However, currently, in the field of MER, the available public datasets have limited sample sizes. Recently, segment-based methods for emotion-related tasks have been proposed, which train backbone networks on shorter segments instead of entire audio clips, thereby naturally augmenting training samples without requiring additional resources. Then, the predicted segment-level results are aggregated to obtain the entire song prediction. The most commonly used method is that the segment inherits the label of the clip containing it, but music emotion is not constant during the whole clip. Doing so will introduce label noise and make the training easy to overfit. To handle the noisy label issue, we propose a semi-supervised self-learning (SSSL) method, which can differentiate between samples with correct and incorrect labels in a self-learning manner, thus effectively utilizing the augmented segment-level data. Experiments on three public emotional datasets demonstrate that the proposed method can achieve better or comparable performance.
Authors: Arnav Kumar Jain, Harley Wiltzer, Jesse Farebrother, Irina Rish, Glen Berseth, Sanjiban Choudhury
Abstract: In inverse reinforcement learning (IRL), an agent seeks to replicate expert demonstrations through interactions with the environment. Traditionally, IRL is treated as an adversarial game, where an adversary searches over reward models, and a learner optimizes the reward through repeated RL procedures. This game-solving approach is both computationally expensive and difficult to stabilize. In this work, we propose a novel approach to IRL by direct policy optimization: exploiting a linear factorization of the return as the inner product of successor features and a reward vector, we design an IRL algorithm by policy gradient descent on the gap between the learner and expert features. Our non-adversarial method does not require learning a reward function and can be solved seamlessly with existing actor-critic RL algorithms. Remarkably, our approach works in state-only settings without expert action labels, a setting which behavior cloning (BC) cannot solve. Empirical results demonstrate that our method learns from as few as a single expert demonstration and achieves improved performance on various control tasks.
Authors: Xiangxiang Gao, Weisheng Xie, Yiwei Xiang, Feng Ji
Abstract: Striking an optimal balance between minimal drafting latency and high speculation accuracy to enhance the inference speed of Large Language Models remains a significant challenge in speculative decoding. In this paper, we introduce Falcon, an innovative semi-autoregressive speculative decoding framework fashioned to augment both the drafter's parallelism and output quality. Falcon incorporates the Coupled Sequential Glancing Distillation technique, which fortifies inter-token dependencies within the same block, leading to increased speculation accuracy. We offer a comprehensive theoretical analysis to illuminate the underlying mechanisms. Additionally, we introduce a Custom-Designed Decoding Tree, which permits the drafter to generate multiple tokens in a single forward pass and accommodates multiple forward passes as needed, thereby boosting the number of drafted tokens and significantly improving the overall acceptance rate. Comprehensive evaluations on benchmark datasets such as MT-Bench, HumanEval, and GSM8K demonstrate Falcon's superior acceleration capabilities. The framework achieves a lossless speedup ratio ranging from 2.91x to 3.51x when tested on the Vicuna and LLaMA2-Chat model series. These results outstrip existing speculative decoding methods for LLMs, including Eagle, Medusa, Lookahead, SPS, and PLD, while maintaining a compact drafter architecture equivalent to merely two Transformer layers.
Authors: Zihao Ye, Lequn Chen, Ruihang Lai, Wuwei Lin, Yineng Zhang, Stephanie Wang, Tianqi Chen, Baris Kasikci, Vinod Grover, Arvind Krishnamurthy, Luis Ceze
Abstract: Transformers, driven by attention mechanisms, form the foundation of large language models (LLMs). As these models scale up, efficient GPU attention kernels become essential for high-throughput and low-latency inference. Diverse LLM applications demand flexible and high-performance attention solutions. We present FlashInfer: a customizable and efficient attention engine for LLM serving. FlashInfer tackles KV-cache storage heterogeneity using block-sparse format and composable formats to optimize memory access and reduce redundancy. It also offers a customizable attention template, enabling adaptation to various settings through Just-In-Time (JIT) compilation. Additionally, FlashInfer's load-balanced scheduling algorithm adjusts to dynamism of user requests while maintaining compatibility with CUDAGraph which requires static configuration. FlashInfer have been integrated into leading LLM serving frameworks like SGLang, vLLM and MLC-Engine. Comprehensive kernel-level and end-to-end evaluations demonstrate FlashInfer's ability to significantly boost kernel performance across diverse inference scenarios: compared to state-of-the-art LLM serving solutions, FlashInfer achieve 29-69% inter-token-latency reduction compared to compiler backends for LLM serving benchmark, 28-30% latency reduction for long-context inference, and 13-17% speedup for LLM serving with parallel generation.
Authors: Dana A Abdullah, Dana Rasul Hamad, Ismail Y. Maolood, Hakem Beitollahi, Aso K. Ameen, Sirwan A. Aula, Abdulhady Abas Abdulla, Mohammed Y. Shakorf, Sabat Salih Muhamad
Abstract: Recognizing the same faces with and without masks is important for ensuring consistent identification in security, access control, and public safety. This capability is crucial in scenarios like law enforcement, healthcare, and surveillance, where accurate recognition must be maintained despite facial occlusion. This research focuses on the challenge of recognizing the same faces with and without masks by employing cosine similarity as the primary technique. With the increased use of masks, traditional facial recognition systems face significant accuracy issues, making it crucial to develop methods that can reliably identify individuals in masked conditions. For that reason, this study proposed Masked-Unmasked Face Matching Model (MUFM). This model employs transfer learning using the Visual Geometry Group (VGG16) model to extract significant facial features, which are subsequently classified utilizing the K-Nearest Neighbors (K-NN) algorithm. The cosine similarity metric is employed to compare masked and unmasked faces of the same individuals. This approach represents a novel contribution, as the task of recognizing the same individual with and without a mask using cosine similarity has not been previously addressed. By integrating these advanced methodologies, the research demonstrates effective identification of individuals despite the presence of masks, addressing a significant limitation in traditional systems. Using data is another essential part of this work, by collecting and preparing an image dataset from three different sources especially some of those data are real provided a comprehensive power of this research. The image dataset used were already collected in three different datasets of masked and unmasked for the same faces.
Authors: Yanbing Zhou, Xiangmou Qu, Chenlong You, Jiyang Zhou, Jingyue Tang, Xin Zheng, Chunmao Cai, Yingbo Wu
Abstract: Prototype-based federated learning has emerged as a promising approach that shares lightweight prototypes to transfer knowledge among clients with data heterogeneity in a model-agnostic manner. However, existing methods often collect prototypes directly from local models, which inevitably introduce inconsistencies into representation learning due to the biased data distributions and differing model architectures among clients. In this paper, we identify that both statistical and model heterogeneity create a vicious cycle of representation inconsistency, classifier divergence, and skewed prototype alignment, which negatively impacts the performance of clients. To break the vicious cycle, we propose a novel framework named Federated Learning via Semantic Anchors (FedSA) to decouple the generation of prototypes from local representation learning. We introduce a novel perspective that uses simple yet effective semantic anchors serving as prototypes to guide local models in learning consistent representations. By incorporating semantic anchors, we further propose anchor-based regularization with margin-enhanced contrastive learning and anchor-based classifier calibration to correct feature extractors and calibrate classifiers across clients, achieving intra-class compactness and inter-class separability of prototypes while ensuring consistent decision boundaries. We then update the semantic anchors with these consistent and discriminative prototypes, which iteratively encourage clients to collaboratively learn a unified data representation with robust generalization. Extensive experiments under both statistical and model heterogeneity settings show that FedSA significantly outperforms existing prototype-based FL methods on various classification tasks.
Authors: Hao Lyu, Yanyong Guo, Pan Liu, Nan Zheng, Ting Wang, Quansheng Yue
Abstract: The use of connected automated vehicle (CAV) is advocated to mitigate traffic oscillations in mixed traffic flow consisting of CAVs and human driven vehicles (HDVs). This study proposes an adaptive deep Koopman predictive control framework (AdapKoopPC) for regulating mixed traffic flow. Firstly, a Koopman theory-based adaptive trajectory prediction deep network (AdapKoopnet) is designed for modeling HDVs car-following behavior. AdapKoopnet enables the representation of HDVs behavior by a linear model in a high-dimensional space. Secondly, the model predictive control is employed to smooth the mixed traffic flow, where the combination of the linear dynamic model of CAVs and linear prediction blocks from AdapKoopnet is embedded as the predictive model into the AdapKoopPC. Finally, the predictive performance of the prosed AdapKoopnet is verified using the HighD naturalistic driving dataset. Furthermore, the control performance of AdapKoopPC is validated by the numerical simulations. Results demonstrate that the AdapKoopnet provides more accuracy HDVs predicted trajectories than the baseline nonlinear models. Moreover, the proposed AdapKoopPC exhibits more effective control performance with less computation cost compared with baselines in mitigating traffic oscillations, especially at the low CAVs penetration rates. The code of proposed AdapKoopPC is open source.
Authors: C\'edric Join, Emmanuel Delaleau, Michel Fliess
Abstract: Model predictive control (MPC) is a popular control engineering practice, but requires a sound knowledge of the model. Model-free predictive control (MFPC), a burning issue today, also related to reinforcement learning (RL) in AI, is reformulated here via a linear differential equation with constant coefficients, thanks to a new perspective on optimal control combined with recent advances in the field of model-free control (MFC). It is replacing Dynamic Programming, the Hamilton-Jacobi-Bellman equation, and Pontryagin's Maximum Principle. The computing burden is low. The implementation is straightforward. Two nonlinear examples, a chemical reactor and a two tank system, are illustrating our approach. A comparison with the HEOL setting, where some expertise of the process model is needed, shows only a slight superiority of the later. A recent identification of the two tank system via a complex ANN architecture might indicate that a full modeling and the corresponding machine learning mechanism are not always necessary neither in control, nor, more generally, in AI.
Authors: Dario Spoljaric, Yashuai Yan, Dongheui Lee
Abstract: Reinforcement-learned locomotion enables legged robots to perform highly dynamic motions but often accompanies time-consuming manual tuning of joint stiffness. This paper introduces a novel control paradigm that integrates variable stiffness into the action space alongside joint positions, enabling grouped stiffness control such as per-joint stiffness (PJS), per-leg stiffness (PLS) and hybrid joint-leg stiffness (HJLS). We show that variable stiffness policies, with grouping in per-leg stiffness (PLS), outperform position-based control in velocity tracking and push recovery. In contrast, HJLS excels in energy efficiency. Despite the fact that our policy is trained on flat floor only, our method showcases robust walking behaviour on diverse outdoor terrains, indicating robust sim-to-real transfer. Our approach simplifies design by eliminating per-joint stiffness tuning while keeping competitive results with various metrics.
Authors: Xingzhi Qian, Xinran Zheng, Yiling He, Shuo Yang, Lorenzo Cavallaro
Abstract: The rapid growth of mobile applications has escalated Android malware threats. Although there are numerous detection methods, they often struggle with evolving attacks, dataset biases, and limited explainability. Large Language Models (LLMs) offer a promising alternative with their zero-shot inference and reasoning capabilities. However, applying LLMs to Android malware detection presents two key challenges: (1)the extensive support code in Android applications, often spanning thousands of classes, exceeds LLMs' context limits and obscures malicious behavior within benign functionality; (2)the structural complexity and interdependencies of Android applications surpass LLMs' sequence-based reasoning, fragmenting code analysis and hindering malicious intent inference. To address these challenges, we propose LAMD, a practical context-driven framework to enable LLM-based Android malware detection. LAMD integrates key context extraction to isolate security-critical code regions and construct program structures, then applies tier-wise code reasoning to analyze application behavior progressively, from low-level instructions to high-level semantics, providing final prediction and explanation. A well-designed factual consistency verification mechanism is equipped to mitigate LLM hallucinations from the first tier. Evaluation in real-world settings demonstrates LAMD's effectiveness over conventional detectors, establishing a feasible basis for LLM-driven malware analysis in dynamic threat landscapes.
Authors: Peng Yan, Ahmed Abdulkadir, Gerrit A. Schatte, Giulia Aguzzi, Joonsu Gha, Nikola Pascher, Matthias Rosenthal, Yunlong Gao, Benjamin F. Grewe, Thilo Stadelmann
Abstract: To go from (passive) process monitoring to active process control, an effective AI system must learn about the behavior of the complex system from very limited training data, forming an ad-hoc digital twin with respect to process inputs and outputs that captures the consequences of actions on the process's world. We propose a novel methodology based on learning world models that disentangles process parameters in the learned latent representation, allowing for fine-grained control. Representation learning is driven by the latent factors influencing the processes through contrastive learning within a joint embedding predictive architecture. This makes changes in representations predictable from changes in inputs and vice versa, facilitating interpretability of key factors responsible for process variations, paving the way for effective control actions to keep the process within operational bounds. The effectiveness of our method is validated on the example of plastic injection molding, demonstrating practical relevance in proposing specific control actions for a notoriously unstable process.
Authors: Florian Pl\"otzky, Katarina Britz, Wolf-Tilo Balke
Abstract: The use of narratives as a means of fusing information from knowledge graphs (KGs) into a coherent line of argumentation has been the subject of recent investigation. Narratives are especially useful in event-centric knowledge graphs in that they provide a means to connect different real-world events and categorize them by well-known narrations. However, specifically for controversial events, a problem in information fusion arises, namely, multiple viewpoints regarding the validity of certain event aspects, e.g., regarding the role a participant takes in an event, may exist. Expressing those viewpoints in KGs is challenging because disputed information provided by different viewpoints may introduce inconsistencies. Hence, most KGs only feature a single view on the contained information, hampering the effectiveness of narrative information access. This paper is an extension of our original work and introduces attributions, i.e., parameterized predicates that allow for the representation of facts that are only valid in a specific viewpoint. For this, we develop a conceptual model that allows for the representation of viewpoint-dependent information. As an extension, we enhance the model by a conception of viewpoint-compatibility. Based on this, we deepen our original deliberations on the model's effects on information fusion and provide additional grounding in the literature.
Authors: Mikel Rodriguez, Raluca Ada Popa, Four Flynn, Lihao Liang, Allan Dafoe, Anna Wang
Abstract: As frontier AI models become more capable, evaluating their potential to enable cyberattacks is crucial for ensuring the safe development of Artificial General Intelligence (AGI). Current cyber evaluation efforts are often ad-hoc, lacking systematic analysis of attack phases and guidance on targeted defenses. This work introduces a novel evaluation framework that addresses these limitations by: (1) examining the end-to-end attack chain, (2) identifying gaps in AI threat evaluation, and (3) helping defenders prioritize targeted mitigations and conduct AI-enabled adversary emulation for red teaming. Our approach adapts existing cyberattack chain frameworks for AI systems. We analyzed over 12,000 real-world instances of AI involvement in cyber incidents, catalogued by Google's Threat Intelligence Group, to curate seven representative attack chain archetypes. Through a bottleneck analysis on these archetypes, we pinpointed phases most susceptible to AI-driven disruption. We then identified and utilized externally developed cybersecurity model evaluations focused on these critical phases. We report on AI's potential to amplify offensive capabilities across specific attack stages, and offer recommendations for prioritizing defenses. We believe this represents the most comprehensive AI cyber risk evaluation framework published to date.
Authors: Shengrui XU, Tianchi Lu, Zikun Wang, Jixiu Zhai, Jingwan Wang
Abstract: Protein-protein interaction (PPI) prediction plays a pivotal role in deciphering cellular functions and disease mechanisms. To address the limitations of traditional experimental methods and existing computational approaches in cross-modal feature fusion and false-negative suppression, we propose SCMPPI-a novel supervised contrastive multimodal framework. By effectively integrating sequence-based features (AAC, DPC, ESMC-CKSAAP) with network topology (Node2Vec embeddings) and incorporating an enhanced contrastive learning strategy with negative sample filtering, SCMPPI achieves superior prediction performance. Extensive experiments on eight benchmark datasets demonstrate its state-of-the-art accuracy(98.13%) and AUC(99.69%), along with excellent cross-species generalization (AUC>99%). Successful applications in CD9 networks, Wnt pathway analysis, and cancer-specific networks further highlight its potential for disease target discovery, establishing SCMPPI as a powerful tool for multimodal biological data analysis.
Authors: Yuhao Wang, Heyang Liu, Ziyang Cheng, Ronghua Wu, Qunshan Gu, Yanfeng Wang, Yu Wang
Abstract: Speech large language models (LLMs) have emerged as a prominent research focus in speech processing. We introduce VocalNet-1B and VocalNet-8B, a series of high-performance, low-latency speech LLMs enabled by a scalable and model-agnostic training framework designed for real-time voice interaction. Central to our contribution is the first application of multi-token prediction (MTP) to speech LLMs. This approach represents a paradigm shift from standard next-token prediction (NTP), offering simultaneous improvements in generation speed and quality. Informed by analysis of MTP's effect on speech generation and experimental comparisons, we designed a straightforward and highly effective MTP implementation. Experiments demonstrate that VocalNet performs on par with mainstream Omni LLMs even with limited training data, and significantly surpasses existing open-source speech LLMs. To foster reproducibility and community advancement, all model weights, inference code, training data, and framework implementations have been made publicly available at https://github.com/SJTU-OmniAgent/VocalNet
Authors: Nirvan Patil, Malhar Abhay Inamdar, Agnivo Gosai, Guruprasad Pathak, Anish Joshi, Aryan Sagavekar, Anish Joshirao, Raj Dandekar, Rajat Dandekar, Sreedath Panat
Abstract: Small Language Models (SLMs) offer efficient alternatives to LLMs for specific domains. The 2023 TinyStories study developed an English dataset that allows SLMs with 1 to 10 million parameters to produce coherent outputs. Our research expands this framework by translating the original dataset into Indian languages and creating synthetic data using LLMs. We focus on Hindi, Marathi, and Bengali, evaluating SLMs for regional language processing and understanding linguistic complexity. We show that SLMs efficiently process regional languages with significantly fewer parameters than LLMs, providing a complementary framework for ``inference based evaluation" of tokenization strategies and linguistic complexity. Our analysis shows that language-specific tokenizers outperform general-purpose ones for Indian languages. Empirical validations, supported by information-theoretic and morphological analyses, provides fundamental understanding behind the better performance of Hindi models over Marathi and Bengali. Additionally, we show that synthetic datasets outperform translated content for training SLMs. Correlation analyses reveal cross-linguistic patterns and language-specific relationships between creativity, grammatical precision, and narrative completeness. These findings advance both the practical application of SLMs to underserved languages and our theoretical understanding of neural language development.
Authors: Zhimin Li, Haichao Miao, Xinyuan Yan, Valerio Pascucci, Matthew Berger, Shusen Liu
Abstract: Recent developments in multimodal large language models (MLLM) have equipped language models to reason about vision and language jointly. This permits MLLMs to both perceive and answer questions about data visualization across a variety of designs and tasks. Applying MLLMs to a broad range of visualization tasks requires us to properly evaluate their capabilities, and the most common way to conduct evaluation is through measuring a model's visualization reasoning capability, analogous to how we would evaluate human understanding of visualizations (e.g., visualization literacy). However, we found that in the context of visualization question answering (VisQA), how an MLLM perceives and reasons about visualizations can be fundamentally different from how humans approach the same problem. During the evaluation, even without visualization, the model could correctly answer a substantial portion of the visualization test questions, regardless of whether any selection options were provided. We hypothesize that the vast amount of knowledge encoded in the language model permits factual recall that supersedes the need to seek information from the visual signal. It raises concerns that the current VisQA evaluation may not fully capture the models' visualization reasoning capabilities. To address this, we propose a comprehensive sanity check framework that integrates a rule-based decision tree and a sanity check table to disentangle the effects of "seeing" (visual processing) and "recall" (reliance on prior knowledge). This validates VisQA datasets for evaluation, highlighting where models are truly "seeing", positively or negatively affected by the factual recall, or relying on inductive biases for question answering. Our study underscores the need for careful consideration in designing future visualization understanding studies when utilizing MLLMs.
Authors: Isabel O. Gallegos, Chen Shani, Weiyan Shi, Federico Bianchi, Izzy Gainsburg, Dan Jurafsky, Robb Willer
Abstract: As generative artificial intelligence (AI) enables the creation and dissemination of information at massive scale and speed, it is increasingly important to understand how people perceive AI-generated content. One prominent policy proposal requires explicitly labeling AI-generated content to increase transparency and encourage critical thinking about the information, but prior research has not yet tested the effects of such labels. To address this gap, we conducted a survey experiment (N=1601) on a diverse sample of Americans, presenting participants with an AI-generated message about several public policies (e.g., allowing colleges to pay student-athletes), randomly assigning whether participants were told the message was generated by (a) an expert AI model, (b) a human policy expert, or (c) no label. We found that messages were generally persuasive, influencing participants' views of the policies by 9.74 percentage points on average. However, while 94.6% of participants assigned to the AI and human label conditions believed the authorship labels, labels had no significant effects on participants' attitude change toward the policies, judgments of message accuracy, nor intentions to share the message with others. These patterns were robust across a variety of participant characteristics, including prior knowledge of the policy, prior experience with AI, political party, education level, or age. Taken together, these results imply that, while authorship labels would likely enhance transparency, they are unlikely to substantially affect the persuasiveness of the labeled content, highlighting the need for alternative strategies to address challenges posed by AI-generated information.
Authors: Rajesh Ranjan, Shailja Gupta, Surya Narayan Singh
Abstract: The rise of autonomous AI agents, capable of perceiving, reasoning, and acting independently, signals a profound shift in how digital ecosystems operate, govern, and evolve. As these agents proliferate beyond centralized infrastructures, they expose foundational gaps in identity, accountability, and ethical alignment. Three critical questions emerge: Identity: Who or what is the agent? Accountability: Can its actions be verified, audited, and trusted? Ethical Consensus: Can autonomous systems reliably align with human values and prevent harmful emergent behaviors? We present the novel LOKA Protocol (Layered Orchestration for Knowledgeful Agents), a unified, systems-level architecture for building ethically governed, interoperable AI agent ecosystems. LOKA introduces a proposed Universal Agent Identity Layer (UAIL) for decentralized, verifiable identity; intent-centric communication protocols for semantic coordination across diverse agents; and a Decentralized Ethical Consensus Protocol (DECP) that could enable agents to make context-aware decisions grounded in shared ethical baselines. Anchored in emerging standards such as Decentralized Identifiers (DIDs), Verifiable Credentials (VCs), and post-quantum cryptography, LOKA proposes a scalable, future-resilient blueprint for multi-agent AI governance. By embedding identity, trust, and ethics into the protocol layer itself, LOKA proposes the foundation for a new era of responsible, transparent, and autonomous AI ecosystems operating across digital and physical domains.
Authors: Guanrou Yang, Chen Yang, Qian Chen, Ziyang Ma, Wenxi Chen, Wen Wang, Tianrui Wang, Yifan Yang, Zhikang Niu, Wenrui Liu, Fan Yu, Zhihao Du, Zhifu Gao, ShiLiang Zhang, Xie Chen
Abstract: Human speech goes beyond the mere transfer of information; it is a profound exchange of emotions and a connection between individuals. While Text-to-Speech (TTS) models have made huge progress, they still face challenges in controlling the emotional expression in the generated speech. In this work, we propose EmoVoice, a novel emotion-controllable TTS model that exploits large language models (LLMs) to enable fine-grained freestyle natural language emotion control, and a phoneme boost variant design that makes the model output phoneme tokens and audio tokens in parallel to enhance content consistency, inspired by chain-of-thought (CoT) and chain-of-modality (CoM) techniques. Besides, we introduce EmoVoice-DB, a high-quality 40-hour English emotion dataset featuring expressive speech and fine-grained emotion labels with natural language descriptions. EmoVoice achieves state-of-the-art performance on the English EmoVoice-DB test set using only synthetic training data, and on the Chinese Secap test set using our in-house data. We further investigate the reliability of existing emotion evaluation metrics and their alignment with human perceptual preferences, and explore using SOTA multimodal LLMs GPT-4o-audio and Gemini to assess emotional speech. Demo samples are available at https://yanghaha0908.github.io/EmoVoice/. Dataset, code, and checkpoints will be released.
Authors: Zhanglin Wu, Tengfei Song, Ning Xie, Weidong Zhang, Mengli Zhu, Shuang Wu, Shiliang Sun, Hao Yang
Abstract: The rapid advancement of large vision-language models (LVLMs) has significantly propelled applications in document understanding, particularly in optical character recognition (OCR) and multilingual translation. However, current evaluations of LVLMs, like the widely used OCRBench, mainly focus on verifying the correctness of their short-text responses and long-text responses with simple layout, while the evaluation of their ability to understand long texts with complex layout design is highly significant but largely overlooked. In this paper, we propose Menu OCR and Translation Benchmark (MOTBench), a specialized evaluation framework emphasizing the pivotal role of menu translation in cross-cultural communication. MOTBench requires LVLMs to accurately recognize and translate each dish, along with its price and unit items on a menu, providing a comprehensive assessment of their visual understanding and language processing capabilities. Our benchmark is comprised of a collection of Chinese and English menus, characterized by intricate layouts, a variety of fonts, and culturally specific elements across different languages, along with precise human annotations. Experiments show that our automatic evaluation results are highly consistent with professional human evaluation. We evaluate a range of publicly available state-of-the-art LVLMs, and through analyzing their output to identify the strengths and weaknesses in their performance, offering valuable insights to guide future advancements in LVLM development. MOTBench is available at https://github.com/gitwzl/MOTBench.
Authors: Nimisha Karnatak, Adrien Baranes, Rob Marchant, Huinan Zeng, Tr\'iona Butler, Kristen Olson
Abstract: Text-based prompting remains the predominant interaction paradigm in generative AI, yet it often introduces friction for novice users such as small business owners (SBOs), who struggle to articulate creative goals in domain-specific contexts like advertising. Through a formative study with six SBOs in the United Kingdom, we identify three key challenges: difficulties in expressing brand intuition through prompts, limited opportunities for fine-grained adjustment and refinement during and after content generation, and the frequent production of generic content that lacks brand specificity. In response, we present ACAI (AI Co-Creation for Advertising and Inspiration), a multimodal generative AI tool designed to support novice designers by moving beyond traditional prompt interfaces. ACAI features a structured input system composed of three panels: Branding, Audience and Goals, and the Inspiration Board. These inputs allow users to convey brand-relevant context and visual preferences. This work contributes to HCI research on generative systems by showing how structured interfaces can foreground user-defined context, improve alignment, and enhance co-creative control in novice creative workflows.
Authors: Xiang Zhang, Yongfeng Zhang
Abstract: The internet is undergoing a historical transformation from the "Internet of Websites" to the "Internet of AgentSites." While traditional Websites served as the foundation for information hosting and dissemination, a new frontier is emerging where AgentSites serve as the hubs of the internet, where each AgentSite hosts one or more AI agents that receive tasks, address them, and deliver actionable solutions, marking a significant shift in the digital landscape and representing the next generation of online ecosystems. Under this vision, AIOS, the AI Agent Operating System, serves as the server for the development, deployment and execution of AI agents, which is a fundamental infrastructure for the Internet of Agentsites. In this paper, we introduce AIOS Server, a runtime framework to host agents and enable global-scale collaboration among decentralized agents. AIOS Server provides a communication protocol leveraging the Model Context Protocol (MCP) and JSON-RPC to enable agent-agent or human-agent interactions. Each AIOS node operates as a server to host and execute agents, while supporting peer-to-peer coordination without reliance on centralized orchestration. Based on AIOS Server, we further present the world's first practically deployed Internet of Agentsites (AIOS-IoA), including AgentHub for agent registration and discovery and AgentChat for interactive communication, at https://planet.aios.foundation. The agent discovery mechanism based on Distributed Hash Tables (DHT) and a Gossip protocol serves as the search engine for the internet of agentsites. This work provides a practical foundation for building the Internet of Agentsites-a new paradigm where autonomous agents become first-class citizens of the web. The implementation is available at https://github.com/agiresearch/AIOS.Server and will be integrated into the AIOS main branch at https://github.com/agiresearch/AIOS.
URLs: https://planet.aios.foundation., https://github.com/agiresearch/AIOS.Server, https://github.com/agiresearch/AIOS.
Authors: Jianhao Yan, Yafu Li, Zican Hu, Zhi Wang, Ganqu Cui, Xiaoye Qu, Yu Cheng, Yue Zhang
Abstract: Recent advances in large reasoning models (LRMs) demonstrate that sophisticated behaviors such as multi-step reasoning and self-reflection can emerge via reinforcement learning (RL) with simple rule-based rewards. However, existing zero-RL approaches are inherently ``on-policy'', limiting learning to a model's own outputs and failing to acquire reasoning abilities beyond its initial capabilities. We introduce LUFFY (Learning to reason Under oFF-policY guidance), a framework that augments zero-RL with off-policy reasoning traces. LUFFY dynamically balances imitation and exploration by combining off-policy demonstrations with on-policy rollouts during training. Notably, we propose policy shaping via regularized importance sampling to avoid superficial and rigid imitation during mixed-policy training. Remarkably, LUFFY achieves an over +7.0 average gain across six math benchmarks and an advantage of over +6.2 points in out-of-distribution tasks. It also substantially surpasses imitation-based supervised fine-tuning (SFT), particularly in generalization. Analysis shows LUFFY not only imitates effectively but also explores beyond demonstrations, offering a scalable path to train generalizable reasoning models with off-policy guidance.
Authors: Shiben Liu, Huijie Fan, Qiang Wang, Baojie Fan, Yandong Tang, Liangqiong Qu
Abstract: Lifelong Person Re-identification (LReID) suffers from a key challenge in preserving old knowledge while adapting to new information. The existing solutions include rehearsal-based and rehearsal-free methods to address this challenge. Rehearsal-based approaches rely on knowledge distillation, continuously accumulating forgetting during the distillation process. Rehearsal-free methods insufficiently learn the distribution of each domain, leading to forgetfulness over time. To solve these issues, we propose a novel Distribution-aware Forgetting Compensation (DAFC) model that explores cross-domain shared representation learning and domain-specific distribution integration without using old exemplars or knowledge distillation. We propose a Text-driven Prompt Aggregation (TPA) that utilizes text features to enrich prompt elements and guide the prompt model to learn fine-grained representations for each instance. This can enhance the differentiation of identity information and establish the foundation for domain distribution awareness. Then, Distribution-based Awareness and Integration (DAI) is designed to capture each domain-specific distribution by a dedicated expert network and adaptively consolidate them into a shared region in high-dimensional space. In this manner, DAI can consolidate and enhance cross-domain shared representation learning while alleviating catastrophic forgetting. Furthermore, we develop a Knowledge Consolidation Mechanism (KCM) that comprises instance-level discrimination and cross-domain consistency alignment strategies to facilitate model adaptive learning of new knowledge from the current domain and promote knowledge consolidation learning between acquired domain-specific distributions, respectively. Experimental results show that our DAFC outperforms state-of-the-art methods. Our code is available at https://github.com/LiuShiBen/DAFC.