cs.AI · 2026-10-15 · No. 144

Artificial Intelligence, 2026-10-15.

34 new papers in cs.AI. Titles, authors, abstracts. Links to arXiv. Want this in your inbox every morning? Subscribe →

01 — The papers

34 entries
  1. 01

    Co-Evolving Acting and Forecasting for Language Agents

    Haofei Yu, Yixing Chen, Yining Zhao, Xuyan Ye, Yuxiang Liu, Zhigang Hua, Yan Xie, Shuang Yang, Jiaxuan You

    cs.AI · cs.LG

    Training language agents is bottlenecked by the high cost of long-horizon rollouts, yet standard agentic reinforcement learning (RL) learns only from a terminal scalar reward, leaving the supervision in intermediate steps largely unused. Prior work exploits this supervision through value targets that estimate per-step success probabilities or observation targets that fit environment response. Both are hard to construct well: (1) value targets...

    arxiv.org/abs/2610.20597 · PDF

  2. 02

    Scaling Discovery Requires Forgetting

    Shangyin Tan, Ben Zhou, Lakshya A Agrawal, Jialin Zhang, Yilin Chen, Qiuyang Mang, Donghyun Lee, Dennis Jacob,...

    cs.AI · cs.CL

    Automated research with LLM agents has shown promising results in various domains. Through iterative refinement and scored verifiers, LLM agents are capable of discovering new ideas. There are different ways to scale automated research: let one agent iterate for longer, run several agents in parallel, or further let the parallel agents communicate and build on one another's candidates. However, we find that all three strategies eventually...

    arxiv.org/abs/2610.20563 · PDF

  3. 03

    PhySGA: A Physics-Space Grounded Agentic Framework for Generalizable and Interactive 3D Composition

    Weichen Lyu, Chong Xia, Mingyu Dong, Long Xu, Shengjun Zhang, Yueqi Duan

    cs.AI · cs.CV

    3D object composition is pivotal to digital content creation and robot simulation, while also serving as a critical testbed for physical intelligence in foundation models. Previous methods either leverage large language models to directly predict object coordinates, or utilize generative diffusion priors to synthesize composite images for visual alignment, both of which rely on static priors in digital spaces. However, these methods...

    arxiv.org/abs/2610.20560 · PDF

  4. 04

    ActiveEvidence: Decision-Focused Evidence Graphs with Verifiable Acquisition Gaps

    Miaobo Hu, Shuhao Hu, Xiaobo Guo, Xin Wang, Bokun Wang, Peng Zhang, Daren Zha, Jun Xiao

    cs.AI · cs.CV · cs.LG

    A multimodal agent must decide whether visual evidence is worth acquiring, which candidate or resolution to query, and when further interaction can no longer change its decision. We study this problem under a strict free-metadata/paid-observation boundary, stochastic vector costs, and a frozen execution trace, and treat it as decision-certifiable active evidence acquisition. A Counterfactual Evidence Interaction Graph (CEIG) represents unary...

    arxiv.org/abs/2610.20532 · PDF

  5. 05

    The Wealth of Agents: Scaling Agentic AI Across Developers

    Austin W. Hanjie, Carlos E. Jimenez, Zirui Wang, Karthik Narasimhan

    cs.AI · cs.MA

    Today's agentic systems are typically designed, built, and deployed by single developers with full control over all participating agents. While this simplifies development, it limits the potential scale of agentic systems. We study multi-developer agentic systems (MDAs) which enable developers to collaboratively build agentic systems to push frontier capability beyond what any individual developer could achieve alone. We construct empirical...

    arxiv.org/abs/2610.20518 · PDF

  6. 06

    SafeAdapt-Poker: Counterfactual Risk Calibration for Budgeted Adaptation under Opponent Shift

    Miaobo Hu, Shuhao Hu, Xiaobo Guo, Xin Wang, Bokun Wang, Rui Chen, Daren Zha, Jun Xiao

    cs.AI · cs.GT · cs.LG

    Opponent histories can mislead after behavioral shifts. SafeAdapt-Poker formulates budgeted adaptation under opponent shift as a dual-counterfactual sequential-control problem using only information available before each decision. At each decision the controller asks two distinct questions: what local downside a deviation from the frozen anchor would induce under matched legal continuations, and what future adaptation value would be lost by...

    arxiv.org/abs/2610.20473 · PDF

  7. 07

    SwarmRSI: Recursive Self-Improvement through Evolving Scientific Agent Swarm

    Qianru Meng, Xi Zhang, Konstantinos Papakostas, Congfeng Cao, Wenxiang Hu, Tianyi Lu, Xiao Zhang, Lun Zhou, Zaiqiao...

    cs.AI · cs.CL

    Recursive self-improvement (RSI) in scientific research aims to improve agents' research capabilities through their own experimental experience. Scientific agent swarms can combine evidence and implementations from different research directions to guide their agents' subsequent experiments. However, poor experimental results can leave unclear whether to revise a hypothesis or its implementation, while overall program scores can hide...

    arxiv.org/abs/2610.20431 · PDF

  8. 08

    From Relevance to Validity: Zero-Shot Tool Selection via Functional and Contextual Validation

    Xiaoxuan Dong, Jin Zhang, Zichen Yuan, Yue Zhou, Zihao Rao, Dexu Yu, Tingyong WU, Wenhao Deng, Junchen Fu, Pengwei...

    cs.AI

    Large language model (LLM) agents increasingly face tasks from unseen domains and execution contexts, making zero-shot tool selection essential for identifying tools that can satisfy task requirements. Existing dominant methods mainly rely on relevance ranking, but semantic matching alone cannot ensure that a tool is functionally compatible or remains applicable in a new context. We preliminarily identify a \emph{transfer-validity gap}, where...

    arxiv.org/abs/2610.20388 · PDF

  9. 09

    Lifted State Policy Optimization: Internalizing Self-Correction through Transition Geometry

    Ren Zhuang, Ben Wang, Shuifa Sun

    cs.AI · cs.LG

    Preference optimization ranks final answers and trajectory RL maximizes aggregate returns, yet neither credits the local change that turns a flawed reasoning state into a better one. We introduce Lifted State Policy Optimization (LSPO), which establishes this finite, cost-adjusted transition as the primary unit of learning. Each textual answer is augmented with a learned continuous coordinate, forming a lifted state that an energy function...

    arxiv.org/abs/2610.20350 · PDF

  10. 10

    TreeFlex: Agents Navigating Their Own Search Trees

    Dongzhuo Li, Fangzhou Cheng, Weiwei Li, Kevin Han, Haixu Ma, Zhongchang Sun, Ruiyi Li, Serena Li, Xiangjun Fan,...

    cs.AI · cs.CL

    A ReAct-style agent accumulates observations, intermediate results, and failed attempts in a single context. As this context grows, stale and off-path information consumes its limited capacity and obscures relevant evidence, making focused reasoning increasingly difficult. We introduce TreeFlex, a framework in which the agent's own policy decides where to search. The agent records each intermediate state, with a short summary, as a node in a...

    arxiv.org/abs/2610.20334 · PDF

  11. 11

    Argument Graph Uncertainty: Quantifying Uncertainty from the Logical Structure of Reasoning Chains

    Simon Baur, Sumire Honda, Wojciech Samek, Jackie Ma

    cs.AI

    Reliable uncertainty estimation is essential for deploying large language models in high-stakes reasoning tasks, yet most existing approaches rely on expensive sampling strategies or shallow token-level signals that ignore the internal logical structure of model outputs. We introduce Argument Graph Uncertainty, a post-hoc framework for multiple choice question tasks, that estimates model confidence directly from the logical structure of a...

    arxiv.org/abs/2610.20328 · PDF

  12. 12

    Reflect After Failure: Brain-Inspired Self-Evolving Agents for Embodied Spatial Intelligence

    Xin Li, Xiaohuan Gu, Yifan Wang, Shaobin Ling, Tianli Shi, Jingao Xu, Zhimeng Yin, Dapeng Wu, Yun Wang

    cs.AI

    Vision-Language Models (VLMs) have made substantial progress in spatial reasoning, and self-evolving agents further extend these capabilities to embodied environments by learning from interactions. However, current self-evolution remains particularly challenging in embodied spatial tasks, complex and multi-step decisions from partial visual evidence making it difficult to determine why an execution fails. The failure may stem from...

    arxiv.org/abs/2610.20312 · PDF

  13. 13

    CASI: Credit Assignment via Set Intervention for Agentic Retrieval

    Xinyi Cheng, Tianfu Zhang, Jiwen Li, Xuesen Zhang, Shihao Yu, Pengfei Zhu, Renhuan, Heda Wang

    cs.AI · cs.IR · cs.LG

    Large language models increasingly generate multiple retrieval actions jointly, while downstream systems often provide only one reward for the resulting set. Broadcasting this reward provides a valid score-function gradient estimator for the expected set-level objective, but does not determine a coalition-dependent within-set attribution. We introduce CASI (Credit Assignment via Set Intervention for Agentic Retrieval), which constructs the...

    arxiv.org/abs/2610.20293 · PDF

  14. 14

    Design Space Exploration of Reinforcement Learning Configurations for Post-Legalization Placement

    Foteini Oikonomou, Maria Pantazi-Kypraiou, George Goudroumanis, Pavlos Stoikos, Athanasios Tziouvaras, Christos...

    cs.AI

    Graph Neural Networks (GNNs) and Reinforcement Learning (RL) have recently been explored for VLSI placement optimization, but their effectiveness for post-legalization placement refinement remains unclear. This paper presents a systematic design space exploration of a GNN-Proximal Policy Optimization (GNN-PPO) framework across three ISPD2015 benchmarks. We evaluate 18 configurations covering reward functions, GNN architectures, action spaces,...

    arxiv.org/abs/2610.20224 · PDF

  15. 15

    Robust Counterfactuals for Markov Decision Processes under Statistical and Causal Uncertainties

    Jessica Lally, Milad Kazemi, Nicola Paoletti

    cs.AI

    Counterfactual reasoning (asking what would have happened, had a decision been different) is a powerful tool for identifying causes of failure and explaining the behaviour of AI systems in sequential decision-making tasks. In Markov Decision Processes (MDPs), two sources of uncertainty complicate counterfactual inference: statistical uncertainty in the MDP transitions when the MDP is estimated from trajectory data, and uncertainty about the...

    arxiv.org/abs/2610.20210 · PDF

  16. 16

    Drivers of LLM Consensus Formation Under Competing Incentives: A Simulation Study

    Isabel Jones, Matthew Edwards, Janina A. Hoffmann

    cs.AI · cs.MA

    Large Language Models (LLMs) are increasingly deployed in multi-agent environments that produce unanimous agreement, both in intended interactions that need coordination and in unintended emergent interactions. Consequently, understanding how LLMs with competing incentives reach consensus is crucial to better interpret and design LLM collectives. We introduce an evaluation, based on the Network Consensus game, to assess the mechanisms driving...

    arxiv.org/abs/2610.20200 · PDF

  17. 17

    HEM-Med: A Hierarchical Agentic Memory Framework for Safe Multi-Center Medication Recommendation

    Min Xiao, Jin Li, Shoujin Wang, Yeguang Zhen, Yishuo Li, Wenpeng Lu

    cs.AI · cs.IR

    Medication recommendation (MR) is critical for clinical decision support. While existing MR methods are mostly designed for single-center learning, where only one independent hospital is involved, multi-center MR with hospitals of different scales has become increasingly demanding in practice. In multi center settings, abundant data from large medical centers can improve recommendation performance for low-resource centers. Despite a few...

    arxiv.org/abs/2610.20191 · PDF

  18. 18

    Value-Guided Repulsion for Offline Reinforcement Learning

    Yixue Hao, Jiahao Zhang, Jiesong Lian, Xufei Guo, Xumeng Guo, Wenhui Zhang, Long Hu, Rui Wang

    cs.AI

    Drifting models have emerged as a highly efficient and stable paradigm for policy design in offline reinforcement learning (RL). However, standard drift policies determine repulsion solely from geometric relationships among generated actions and therefore cannot distinguish them by expected value. To address this limitation, we introduce Value-Guided Repulsion Q-Learning (VRQL) to integrate action value estimates into the particle interaction...

    arxiv.org/abs/2610.20190 · PDF

  19. 19

    LLM-Based Policy Compliance Verification of Loan Applications

    Florentina Voboril

    cs.AI · cs.CL

    Banks rely on extensive internal policy frameworks to ensure that corporate lending decisions comply with regulatory requirements and internal risk standards. Reviewing loan applications against these policies is a largely manual and knowledge-intensive process, which can lead to overlooked policy violations or exceptions. This report presents a prototype of an LLM-based policy checker that automatically verifies loan applications against...

    arxiv.org/abs/2610.20155 · PDF

  20. 20

    Maximising Transparency in Context Compaction Improves Retention of Facts and Commitments for Lightweight Coding Agents

    Felix Woolford

    cs.AI

    Long coding agent sessions outgrow the model's context window, so agents compact: older history is replaced by something shorter, usually a model-written summary. A summary selects and rewords, and the agent cannot see what was lost. Existing work has explored a range of compaction techniques, but comparatively little attention has been paid to their effectiveness with lightweight models, including those deployed locally or used for...

    arxiv.org/abs/2610.20116 · PDF

  21. 21

    EvalBuilder: Can LLMs Build Efficient Benchmarks?

    Huaisheng Zhu, Hangfan Zhang, Teng Xiao, Siyuan Xu, Shijie Zhou, Zhimeng Guo, Zhengyu Chen, Jingang Wang

    cs.AI

    Large language model (LLM) agents are increasingly used for long-horizon research and engineering, yet evaluating these capabilities remains expensive, time-consuming, and heavily dependent on human expertise. This motivates a natural question: can LLM agents themselves construct efficient benchmarks that preserve the conclusions of costly reference evaluations? We study this capability by formulating automatic benchmark development as a...

    arxiv.org/abs/2610.20099 · PDF

  22. 22

    Designing Textual Interfaces To Enhance Llm Reasoning On Audio-Visual Question Answering

    Yi Wang, Tianzi Wang, Tong Lei, Dahan Wang, Peter Bell

    cs.AI · cs.CL · eess.AS

    Automatic prompt optimization (APO) has shown strong effectiveness for text-based LLMs and LLM systems, but its application to audio-visual understanding systems requires suitable textual interfaces through which multimodal reasoning can be exposed to text-based optimizers. We investigate how the design of textual interfaces affects prompt optimization for multimodal reasoning, on the audio-visual question answering task. We compare three...

    arxiv.org/abs/2610.20085 · PDF

  23. 23

    Beyond Correctness: Compute-Responsive Data Selection for Reasoning Model Post-Training

    Liang He, Xiangyu Li, Yuxin Wang, Hanbin Zhang, Mingyu Cao, Zhaomiao Guo, Xilu Wang

    cs.AI

    Reasoning language models increasingly rely on test-time compute scaling, yet most existing work studies how to allocate inference computation while treating the model itself as fixed. We ask a complementary question: can reasoning models be post-trained to make better use of a given amount of test-time compute, and which examples are most useful for this goal? Our key observation is that reasoning problems differ substantially in their...

    arxiv.org/abs/2610.20057 · PDF

  24. 24

    Memorizing Who, Knowing How: Benchmarking Personal and Procedural Memory for Persistent Assistants

    Maosheng Qin, Qi Zhang, Mingxuan Xia, Jinglei Chen, Jiexiang Wang, Sheng Guo, Cheng Peng, Haobo Wang

    cs.AI

    Memory systems for persistent personal agents must move beyond isolated memories, maintaining both personal memory about whom they serve and procedural memory about how to act. Crucially, in harness-based personal assistance, procedures or skills are inherently user-conditioned: personal context determines which procedures apply, how they are instantiated, and when they should change. Existing benchmarks typically emphasize either personal or...

    arxiv.org/abs/2610.20022 · PDF

  25. 25

    What Coding Agents Leave Unsaid: Measuring Task-State Gaps Between Execution and Reports

    Guoyang Xu, Rui Zhang, Xiuyu Wu, Hao Chen

    cs.AI

    Human supervisors rely on coding agents' reports to assess progress, verify repairs, and decide when to intervene. But a supervisor cannot investigate a failure they were never told occurred. Failed attempts, workarounds, and unresolved risks may disappear from routine reports, leaving supervisors unaware of issues that warrant scrutiny. We conduct CTXGAP, a measurement study quantifying the gap in task-state recoverability between agents'...

    arxiv.org/abs/2610.20003 · PDF

  26. 26

    Neurosymbolic Learning of Interpretable World Models

    Riccardo Andreoni, Massimiliano Ronzani, Chiara Ghidini, Alessandro Daniele, Emile van Krieken

    cs.AI

    World models aim to learn compact representations of an environment from observations, enabling agents to predict future states and plan their behaviour. However, most existing approaches learn black-box representations, with limited ability to incorporate prior knowledge about the world. We introduce Concept-Grounded World Models (CGWMs), a neurosymbolic world model design that instead learns interpretable concept representations. We learn...

    arxiv.org/abs/2610.19955 · PDF

  27. 27

    When Protection Overrides Consequences: Understanding Protection Mode in Large Language Model Decision-Making

    Yuanfan Li, Junyi Chen, Qi Zhou, Rui Yang, ZhangLiujie, Weihang Chen

    cs.AI

    Frontier language models excel at complex tasks yet may prioritize protection over larger physical consequences in simple, urgent decisions. GPT-5.6 Sol, for example, moves a girl playing in shallow water before retrieving a submerged camera. We call this Protection Mode and study it with ProtectionBench: 285 prompts, 34 physical families, and nine models. At lowest effort, at least one of four responses prioritizes the subject on 34.74% of...

    arxiv.org/abs/2610.19948 · PDF

  28. 28

    Sparse Transformer: a Controlled Testbed for Mechanistic Interpretability

    Anton Korznikov, Alexey Dontsov, Elena Tutubalina

    cs.AI

    Mechanistic interpretability aims to decompose neural network computations into human-understandable features, yet progress is limited by the absence of ground-truth units to evaluate against. We address this with a Sparse Transformer (ST): a 5-layer language model trained on TinyStories whose wide, TopK-sparse MLP forces neuron specialization. We show that ST neurons are highly interpretable: they outperform a loss-matched dense transformer...

    arxiv.org/abs/2610.19907 · PDF

  29. 29

    How GUI Agents Fail: GraceBench for Execution Stability Diagnosis

    Ridong Zhang, Shuxian Liang, Chaoqun Wan, Zhen Huang, Xian-Sheng Hua

    cs.AI

    As GUI agents are increasingly deployed for mobile task automation, their reliability depends not only on final task outcomes, but also on whether they reach these outcomes through stable interaction trajectories. In automation workflows, unnecessary actions, detours, and execution errors can add noise to collected data, introduce irreversible side effects, or undermine trust, even when the agent eventually completes the task. It is therefore...

    arxiv.org/abs/2610.19900 · PDF

  30. 30

    Explorer-Zero: Improving Self-Play for Long-Horizon Tasks with Decomposable Verifiable Rewards

    Tianyu Liu, Davide Belli, Bence Major, Arash Behboodi, Tribhuvanesh Orekondy

    cs.AI · cs.LG

    Training LLM agents on long-horizon interactive tasks with reinforcement learning typically relies on human-annotated tasks and reliable verification, both expensive to construct for stateful environments. Self-play offers an alternative, but must generate achievable, compositional tasks and provide feedback beyond binary outcomes. We propose Explorer-Zero, a self-play framework that jointly generates tasks and fine-grained, verifiable...

    arxiv.org/abs/2610.19868 · PDF

  31. 31

    Euler: Towards Human-Centered Multi-Agent Mathematical Discovery

    Qiuyu Chen, Qi Wang, Liang Xu, Xin Jin, Baao Xie, Wenjun Zeng

    cs.AI

    Mathematical research drives scientific discovery, and recent advances in AI for mathematics have shown that language-model agents can solve challenging problems and develop substantial mathematical proofs. Yet much of today's AI-for-mathematics work centers on obtaining a final answer. This emphasis leaves the exploration behind a result difficult for researchers to understand, while long-running investigations face fragmented context and...

    arxiv.org/abs/2610.19864 · PDF

  32. 32

    LAGO: Language Agents with Reusable Growing Options

    Yuran Wu, Mengkang Hu, Qionglin Qiu, Fanqing Meng, Qiguang Chen, Yuhang Zhou, Qiudan Zhang, Xu Wang

    cs.AI · cs.MA

    Long-horizon tasks expose a weakness of monolithic tool-use loops: one agent must decompose, execute, and verify a task within a single trajectory. We introduce LAGO, which treats specialized language agents as reusable temporally extended options. A meta-agent retrieves options from a persistent library or spawns new ones, dispatches sub-tasks, verifies results, and admits successful novel options. The library grows across tasks without...

    arxiv.org/abs/2610.19843 · PDF

  33. 33

    Teaching Agents in their Own Style: Blackbox Teacher Distillation for Web Agents

    Md Farhan Ishmam, Kenneth Marino

    cs.AI

    Frontier web agents using closed models have high inference and computational costs. Smaller open models can be an alternative, but to be performant, they require training supervision from a human or teacher model. Token-level teacher distillation is not an option for closed teachers as it needs access to the model's logits. The standard approach is to imitate teacher demonstrations offline, but it trains the student on states it has never...

    arxiv.org/abs/2610.19834 · PDF

  34. 34

    ActionOPD: Distilling Decisions, Not Tokens

    Huining Yuan, Xiangmin Yi, Zelai Xu, Zixu Wang, Xinlei Chen, Xiao-Ping Zhang, Yu Wang, Chao Yu

    cs.AI

    Distilling capable language-model agents into smaller models requires aligning the student's behavior with the teacher's. This raises a fundamental question: at which states and at what granularity should this alignment be performed? Offline imitation provides dense but token-specific supervision and can suffer from state-distribution mismatch. Token-level on-policy distillation (OPD) addresses this mismatch but typically requires teacher...

    arxiv.org/abs/2610.19820 · PDF

This edition is part of The Daily Abstract — cs.AI archive. Subscribe to receive these in your inbox each morning, automatically translated to Spanish, with reply-to-PDF: arxivdaily.ignorelist.com.

Colophon Set in Georgia, with system sans for interface chrome and a monospaced stack for code and paper identifiers. Sole accent: amber #D99C5E. Built and served on an always-free VM. The masthead is set 14% letterspaced because newspapers do that and it works.