cs.AI · 2026-08-25 · No. 95

Artificial Intelligence, 2026-08-25.

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

01 — The papers

56 entries
  1. 01

    ReWorld: An Interactive World Model with Long-Horizon Memory

    Zhifei Chen, Luozhou Wang, Guibao Shen, Dongyu Yan, Shuai Yang, Tianshuo Xu, Yihua Du, Wei Wang, Tianyi Gui,...

    cs.AI

    An interactive world model must follow the user's actions, remember the places it has shown, and stream in real time. The tension is structural: control wants a short horizon, memory wants an unbounded one. ReWorld separates the two during training and bounds them at inference. Mixed per-head attention windows confine most heads to the recent past while a small set of global heads attends over the entire history, and random head routing keeps...

    arxiv.org/abs/2608.23565 · PDF

  2. 02

    Prime Agent: A Self-Improving RLM Harness

    Seth Karten, Alex L. Zhang, Kevin Thomas, Sebastian Müller, Elie Bakouch, Daniel Auras, Mika Senghaas, Fares Obeid,...

    cs.AI · cs.CL · cs.SE

    Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context. Prime Agent is an open-source harness for long-horizon evaluation and coding-agent workflows. A persistent IPython REPL follows the Recursive Language Model abstraction for programmatic context processing and test-time compute, while Continual Harness preserves histories, memories, skills,...

    arxiv.org/abs/2608.23552 · PDF

  3. 03

    How AI Assistance Affects Human Skill Development: A Study of Learning with Logic Puzzles

    Shang Wu, Catarina G Belem, Shuyuan Fu, Mark Steyvers, Padhraic Smyth

    cs.AI

    While AI assistance can improve human task performance in the short term, it may also undermine the development of skills in the longer term. We examine this tension in a controlled logic-puzzle experiment involving on-demand AI assistance, where participants complete tasks before, during, and after AI is available. By experimentally varying AI request costs, we find that lower-cost assistance induces more frequent AI use. We also find that...

    arxiv.org/abs/2608.23543 · PDF

  4. 04

    Correcting a learned physical invariant improves world-model rollouts

    Richard Bao

    cs.AI

    World models can predict video without learning dynamics that they reliably preserve. We test whether a frozen DreamerV3 trained only on pendulum video learns a scalar that its own latent transition treats as approximately conserved. A label-free search recovers the same energy-like invariant across independently trained conservative models, while the same procedure finds no comparable invariant in matched damped models. During autonomous...

    arxiv.org/abs/2608.23526 · PDF

  5. 05

    EarthVerse: Benchmarking Scientific Agents Across Dynamic Earth Systems and Natural Hazards

    Zhiqing Cui, Xinxiang Yin, Yihong Tang, Xinglang Zhang, Yuanzhe Hu, Siru Zhong, Weidong Tang, Yuxuan Liang, Weijia...

    cs.AI

    Earth-system analysis reconstructs changing physical processes from observations that differ in source, scale, timing, and modality. Natural hazards make this work consequential because incomplete evidence can change estimates of severity, exposure, and mechanism. We introduce EarthVerse, a benchmark that evaluates scientific agents through package-scoped investigations. Its 405 reproducible tasks are grounded in 199 documented events and 19...

    arxiv.org/abs/2608.23525 · PDF

  6. 06

    Mitigating Reasoning-Induced Misalignment via Safety-Direction Penalty

    Yipeng Zhao, Qishun Yang, Shenzhe Zhu, Shu Yang, Di Wang

    cs.AI · cs.CL

    Reasoning-Induced Misalignment, where fine-tuning on reasoning data containing no harmful content, including mathematics, code, and problem-solving with chain-of-thought traces can induce harmful behaviors of LLM, posing a serious challenge to the safety of LLM reasoning. Cross-architecture, cross-scale, and cross-dataset checks show that RIM does not always emerge. Previous work attributed RIM to neuron-level entanglement, but did not...

    arxiv.org/abs/2608.23497 · PDF

  7. 07

    SRPO: Self-Reflective Policy Optimization for Long-Horizon Reasoning

    Jialong Liu, Yuling Shi, Ning Yang, Xiaodong Gu, Zuchao Li

    cs.AI

    Self-reflection is a powerful mechanism for credit assignment in human learning, converting sparse outcome feedback into actionable guidance. However, its potential for post-training Large Language Models (LLMs) remains underexplored. We propose Self-Reflective Policy Optimization (SRPO), a framework that internalizes this capability. SRPO enables LLMs to analyze their own completed trajectories, synthesize errors into concise "reflection...

    arxiv.org/abs/2608.23493 · PDF

  8. 08

    Multi-Modal Semantic Expansion with Constrained LLM Reranking for Conversational Music Recommendation

    Naman Garg, Sarika Jain, George Fazekas

    cs.AI

    We present Team Semiintelligencn's solution for the ACM RecSys 2026 TalkPlayData Challenge, addressing conversational music recommendation through a multi-modal and personalized conversational recommender system. Our submitted system employs a three-stage pipeline: (1) multi-modal retrieval constructing decay-weighted centroids across seven dense embedding spaces - track- and user-level CF-BPR, Qwen3 (metadata, lyrics, attributes), CLAP...

    arxiv.org/abs/2608.23484 · PDF

  9. 09

    StrategyBench: Evaluating Explicit Strategy Induction in Large Language Models

    Jinghan Tan, Yuanzheng Wang, Lu Chen, Zijun Chen, Yuqian Wang, Maosong Sun

    cs.AI

    As large language models are increasingly used in data-scarce and evolving task scenarios, few-shot in-context learning (ICL) has become a key paradigm for task adaptation. However, direct ICL often uses a small set of examples without explicitly abstracting task rules, making it sensitive to example construction. In contrast, human learners often reduce such sensitivity by first summarizing task rules from examples and then applying them to...

    arxiv.org/abs/2608.23475 · PDF

  10. 10

    Characterizing Necessary Losers to Explain Tournaments Losers

    Contet Clément, Umberto Grandi, Jérôme Mengin

    cs.AI

    We study the problem of formally explaining why a candidate was not selected by a given tournament rule, by identifying sub-tournaments in which the candidate loses independently of how the rest of the tournament is completed. We define destructive minimal supports as any minimal sub-tournaments satisfying this property, which in formal explainable artificial intelligence correspond to abductive explanations for the question "Why does the...

    arxiv.org/abs/2608.23446 · PDF

  11. 11

    SkillAlchemy: Open-World Agent Skill Creation

    Hengjun Wang, Shuyue Wei, Boyi Liu, Jun Yang, Yongxin Tong

    cs.AI

    Agent skills are reusable procedural artifacts that extend language agents with specialized workflows, tool conventions, and domain behaviors at inference time. However, creating reliable skills still depends largely on human authorship, model priors, or execution traces. These sources are often unavailable for unfamiliar tasks, suggesting the need to create skills from open-world materials. In this paper, we study open-world skill creation:...

    arxiv.org/abs/2608.23417 · PDF

  12. 12

    MediSkill-Evo: Process-Constrained Self-Evolution for Evidence-Grounded Clinical Interaction

    Ruoyu Wu, Shenfu Xie, Yinqian Sun, Haibo Tong, Feifei Zhao

    cs.AI

    Interactive clinical agents must gather decisive evidence and convert it into grounded actions under partial observability. A correct final diagnosis alone does not show that an agent respected evidence and care-process constraints. We introduce MediSkill-Evo, a clinical agent that evolves governed process knowledge without backbone fine-tuning. It separates experience into four typed banks for clinical skills, process rules, symbolic...

    arxiv.org/abs/2608.23397 · PDF

  13. 13

    Modalities Should Talk to Each Other: Dual-Stream Multimodal Learning for Long-Horizon Influenza Forecasting

    Seyed Mohammad Hossein Hashemi, Mohsen Hooshmand, Parvin Razzaghi

    cs.AI · stat.AP

    Forecasting long-range influenza-like illness (ILI) matters for public health readiness. Publicly available surveillance datasets typically pair numeric epidemiological signals with textual information that is noisy, loosely structured, only indirectly related to near-term trends, and often lagged relative to the numeric signal. Fusing the two therefore requires careful design. We propose Dual-Stream Attention (DSA), a multimodal deep...

    arxiv.org/abs/2608.23373 · PDF

  14. 14

    Walking on the DARKSIDE

    Aldo Gangemi, Emanuele Bottazzi

    cs.AI · cs.LO

    Large Language Models (LLMs) recognise patterns but do not natively track the path of exclusions that a coherent discourse demands. When an input rests on a fabricated authority, a misapplied mechanism, or a surreptitious analogy, an unsteered LLM tends to engage with it as if it were grounded, and to reify the misstep into any structured output it generates. Logic-Augmented Generation (LAG) with POLANYI++, an LLM-steering method that uses...

    arxiv.org/abs/2608.23370 · PDF

  15. 15

    Agent-G$^2$: Gaussian Guidance for Agentic Reinforcement Learning

    Zixuan Wang, Yanrui Miao, Zhengxi Lu, Teng Pan, Yiwen Qiu, Hongxing Li, Peng Qiu, Ruiqing Zhang, Yongliang Shen

    cs.AI · cs.CL

    Hint-based reinforcement learning addresses reward sparsity in long-horizon agentic tasks by retaining a prefix of an expert trajectory before each rollout, letting the policy explore from a state closer to success. Its effectiveness hinges on the guidance depth: how much of the trajectory to keep. Existing methods treat this depth as a deterministic scalar. Scheduled approaches share one value across samples and ignore per-task...

    arxiv.org/abs/2608.23318 · PDF

  16. 16

    EviSafe: Evidence-Grounded Safety Evaluation for Vision-Language Models

    Xuetong Li, Gaofeng Liu

    cs.AI

    Vision-language model safety benchmarks typically evaluate only final responses: whether a model refuses, warns, or complies. This outcome-level view cannot tell whether a model is safe for the right multimodal reason. Safelooking behavior may reflect keyword-triggered refusal, missed visual hazards, or over-refusal of benign-sensitive inputs. We introduce EviSafe, an evidence-grounded framework for VLM safety that jointly evaluates natural...

    arxiv.org/abs/2608.23313 · PDF

  17. 17

    Apodex 1.1: Scaling Agentic Intelligence for Complex Work

    Apodex Team, B. An, B. Li, B. Wang, B. Zhang, B. L. Wang, C. Feng, C. Wei, C. Xue, C. Zhang, D. Ng, D. Ye, E. Min,...

    cs.AI · cs.CL · cs.LG

    General-purpose language models can reason and synthesize knowledge, but complex work also requires sustained interaction with files, information sources, and executable code, together with state maintenance, failure recovery, and verifiable delivery. We call this \emph{working capability}: sustained, verifiable progress toward a real-world objective. Apodex 1.1 develops this capability along two complementary dimensions. \emph{Environment...

    arxiv.org/abs/2608.23283 · PDF

  18. 18

    Hidden in the Request: Explaining Unethical LLM Compliance through Token Relevance

    Or Biton, Tomer Krichli, Itai Allouche, Joseph Keshet

    cs.AI · cs.CL

    Although Large Language Models (LLMs) are aligned to optimize for both helpfulness and harmlessness, these dual objectives may conflict, inevitably leading to alignment failures. This work systematically investigates instances where LLMs fail to exhibit ethical behavior. To understand the underlying mechanics of these vulnerabilities, we introduce a probing methodology that presents unethical scenarios to LLMs in three distinct structural...

    arxiv.org/abs/2608.23264 · PDF

  19. 19

    Automated Construction of FAIR Digital Object Knowledge Graphs from Flat Cultural Heritage Records

    Zeyd Boukhers, Lingxiao Kong, Xenophon Zabulis, Georgios Toubekis

    cs.AI · cs.CL · cs.DL

    The FAIR Digital Object (FDO) framework mandates that metadata attribute values be expressed as persistent identifiers (PIDs) wherever possible, to produce a fully machine-actionable graph in which every reference is resolvable. The Europeana Data Model was designed long before the FDO specification, and it stores most metadata values as plain text. This serves human browsing well enough, but gives an automated agent nothing to follow across...

    arxiv.org/abs/2608.23263 · PDF

  20. 20

    Is Next-Chunk Reasoning RL Really Better than SFT? Revisiting Training Strategies under no-CoT Data

    Yinhao Tang, Youqing Fang, Yanan Sun, Jiangning Liu, Ziyi Wang, Xun Zhao, Weiming Zhang, Bin Liu, Kuikun Liu, Wenwei...

    cs.AI

    Recent work proposes next-chunk reasoning RL for leveraging no-CoT data---corpora such as worked solutions and textbook derivations that contain reasoning-rich content but lack explicit chain-of-thought annotations. The method trains a model to generate implicit reasoning traces and rewards them by their ability to predict the next chunk of text. While promising, existing evaluations primarily compare against conventional SFT baselines,...

    arxiv.org/abs/2608.23256 · PDF

  21. 21

    What is mathematics now, and what should it be?

    Jeremy Avigad

    cs.AI · math.HO

    Advances in neural theorem provers have been impressive, but the successes obscure a broader vision of what AI can do for mathematics and how mathematicians can engage with AI. This essay advances a more expansive and optimistic point of view.

    arxiv.org/abs/2608.23218 · PDF

  22. 22

    Cognitive Profiling of LRMs' Reasoning Traces Using Bloom's Taxonomy

    Maria-Eleni Zoumpoulidi, Georgios Paraskevopoulos, Alexandros Potamianos

    cs.AI · cs.CL

    Large Reasoning Models (LRMs) have revolutionized reasoning in LLMs, and the increasing public availability of reasoning traces creates valuable opportunities to study model behavior not only at the surface level but also at the granularity of individual reasoning steps. However, understanding the types of thinking employed during reasoning - which offers critical insights into models' reasoning patterns and enables actionable applications -...

    arxiv.org/abs/2608.23205 · PDF

  23. 23

    AI emotional support is better only when chosen, but shifts preferences even when it is not

    Yaoxi Shi, Cathy Mengying Fang, Guy LabanPattie Maes, Amit Goldenberg

    cs.AI · cs.HC

    People increasingly face a novel decision when seeking emotional support: human or AI. In existing studies, AI's empathic messages are rated as well as or better than humans'. But these studies either assigned the support source or honored people's choice. In real life, support is often incongruent with choice, as people want one source and receive the other. Across three experiments (N = 1,951), participants chose whether to share an...

    arxiv.org/abs/2608.23196 · PDF

  24. 24

    Jiuge-Tuiqiao: An Interpretable Human-AI System for Classical Chinese Poetry Refinement

    Yufeng Han, Lifan Deng, Cunliang Kong, Wenhao Li, Xin Cong, Yuzhuo Bai, Kangyang Luo, Maosong Sun

    cs.AI

    Classical Chinese poetry composition has long valued Tuiqiao, the iterative refinement of words, imagery, and prosody. However, many current AI poetry systems follow a one-shot generation paradigm, which reduces users to prompt providers and weakens their creative agency. We present Jiuge-Tuiqiao, an interactive human-AI collaborative system for classical Chinese poetry composition. The system is designed around a triadic model: user-driven...

    arxiv.org/abs/2608.23098 · PDF

  25. 25

    POOL: Propagated Uncertainty Over Lookalikes

    Rounak Sharma, Ananya B. Sai, Soumyabrata Pal

    cs.AI

    Black-box large language models need confidence scores that can separate likely-correct from likely-incorrect outputs, enabling systems to prioritize human review, route uncertain cases to stronger models, or choose abstention thresholds on development data. Yet existing confidence estimators face a cost-quality trade-off: verbal confidence is cheap but is often overconfident, while sampling-based uncertainty is more informative but scales...

    arxiv.org/abs/2608.23086 · PDF

  26. 26

    AgentWeave: Routing Before Reasoning for Efficient Function Calling in Tool-Rich Language Models

    Saurav Singla, Aarav Singla, Advik Gupta, Parnika Gupta

    cs.AI · cs.CL

    Large language models increasingly operate over large collections of tools, functions, APIs, and specialized agents. As the candidate action space grows, a function-calling model must process more schemas, consume more prompt tokens, and distinguish among increasingly similar or irrelevant alternatives. We study a complementary systems strategy: reduce the candidate set before language-model inference while leaving the downstream model...

    arxiv.org/abs/2608.23078 · PDF

  27. 27

    From Generation to Simulation: How Far Are World Models from Being True Simulators?

    Tong Wang, Huan Deng, Mucheng Yang, Yang He, Xiaohui Kuang, Gang Zhao

    cs.AI · cs.CV

    With the rapid progress of diffusion models and large-scale video generation, generative world models are increasingly expected to replace traditional simulators, including physics engines, game engines, and reinforcement-learning environments. Yet the remaining distance from generation to simulation lacks a systematic assessment. We present a capability-based study using an external yardstick: eight capabilities of a traditional simulator,...

    arxiv.org/abs/2608.23070 · PDF

  28. 28

    Improving O-RADS Risk Stratification from Ultrasound Reports: A Comparative Evaluation of Hybrid versus End-to-End LLM Reasoning Strategies

    Xiaotong Tan, Chunli Qiu, Xin Liu, Qing Huang, Guangli Zhou, Bo Gao, Xiaoyan Song, Shuyan Wang, Xiuqin Wang, Wufeng...

    cs.AI

    Background: Automating clinical guideline-based decision-making with large language models (LLMs) remains challenging because of reliability, hallucination, and limited interpretability. We compared the performance of LLMs and reasoning strategies for automated Ovarian-Adnexal Reporting and Data System (O-RADS) classification from free-text pelvic ultrasound reports. Methods: In this retrospective study, consecutive patients with ovarian...

    arxiv.org/abs/2608.23061 · PDF

  29. 29

    LLM-based Agents for Forecasting and Prediction: Methods, Training, Evaluation, and Applications

    Xiaogang Xu, Jiaqi Tang, Jianmin Chen, Yingying Yan, Zhenchao Tang, Xiangxin Zhou, Xiaobin Hu, Wei Wei, Jinfeng Wu,...

    cs.AI

    Large language models (LLMs) now support forecasting systems that combine language-based reasoning with temporal data, evidence retrieval, external tools, and iterative prediction. We investigate LLM-based forecasting agents, meaning systems in which a language model contributes to a scored prediction about a future or currently unobserved target. We organize architectures into three groups. Standalone LLM workflows operate on encoded time...

    arxiv.org/abs/2608.23058 · PDF

  30. 30

    From Inertia to Objectivity: Improving Deep Research Agents with Noise Isolation

    Xiangxin Zhang, Zhanwei Zhang, Zhihang Fu, Binbin Lin, Wenxiao Wang

    cs.AI

    Web search agents powered by Large Language Models (LLMs) show strong promise, but deep research tasks expose a recurring failure mode: once an agent has produced a query, plan, or intermediate conclusion, it becomes less objective when later judging the consequences of that same action. We term this phenomenon \textbf{inertia bias}. To make it measurable, we introduce the IBIS benchmark, which controls the search observations while varying...

    arxiv.org/abs/2608.23045 · PDF

  31. 31

    AutoSaddler: Automatic Harness Optimization with Durable Updates from Agent Execution Traces

    Sungho Park, Wonjoong Kim, Rongyuan Tan, Jue Zhang, Wook-Shin Han, Pengfei Gao, Chanyoung Park, Yongqiang Yao, Rao...

    cs.AI · cs.CL · cs.LG · cs.MA · cs.SE

    LLM agents remain unreliable on long-horizon tasks, where small local failures can compound over extended interactions and lead to overall task failure. Although external harnesses can substantially improve robustness, harness design remains a manual and expensive process that requires searching over a large space of prompts, tool configurations, and control logic. We propose AutoSaddler, an automatic harness optimization framework that...

    arxiv.org/abs/2608.23041 · PDF

  32. 32

    MobilePA-Bench: Benchmarking Mobile Planner Agents on Complex Real-World Tasks

    Yi Zhu, Xiongwei Wu, Qiyi Wang, Tingyu Qu, Jiajun Liu, Sihan Cao, Long Chen, Weigao Sun, Feida Zhu, Yiran Zhong, Steven Hoi

    cs.AI

    As on-device LLM agents evolve into personal copilots, the mobile operating system has become a key testbed for this paradigm, making rigorous capability evaluation essential. Yet existing benchmarks fall into two camps, each with a critical blind spot: GUI-centric benchmarks test surface-level screen manipulation while overlooking background tool use and long-horizon planning, whereas static function-calling benchmarks rely on offline API...

    arxiv.org/abs/2608.23035 · PDF

  33. 33

    Artificial Empathy: Towards a Framework for Unsupervised Agency Detection and Policy Reconstruction

    Peter Kuhn, Chris Pang, Sonakshi Chauhan

    cs.AI

    We study how an AI system can identify and model other agents in its environment from observation alone, which is a capability necessary for cooperative behaviour in the real world. This problem is less constrained than inverse reinforcement learning and remains largely unexplored. We propose a framework that uses a reinforcement learning agent, trained on an independent task as a prior about agentic dynamics, to perform agency detection and...

    arxiv.org/abs/2608.23030 · PDF

  34. 34

    PsychJail: Exploring Psychological Jailbreaks via Multi-Turn Persuasion of LLM Policies

    Zeyu Feng, Qingyu Wu, Yuzhe Luo, Hua Cheng

    cs.AI

    Large language models (LLMs) are increasingly deployed in education, healthcare, policy advising, and other interactive settings, where users engage them as sustained social interlocutors rather than one-shot query engines. This shift makes jailbreaks a growing safety threat, yet most research emphasizes single-turn prompt optimization or iterative attack refinement, leaving psychologically grounded multi-turn vulnerabilities underexplored....

    arxiv.org/abs/2608.23028 · PDF

  35. 35

    PatchWrite: One Line, Not One Section -- Compile-Gated, Validity-Preserving Editing for AI-Drafted Manuscripts

    Weiwei Yang

    cs.AI · cs.CL · cs.SE

    Automated manuscript pipelines often regenerate an entire section to repair a local defect, allowing unrelated metrics and citations to change even when the resulting PDF still builds. PatchWrite instead constrains how candidate edits become committed manuscript states: it reuses bounded EDIT N M editing and rollback, but tightens compilation acceptance with fatal-log checks and adds evidence locks that require every cited key and...

    arxiv.org/abs/2608.23001 · PDF

  36. 36

    SA-RSQ: A Versatile Sparse Representation Framework for Multi-modal Recommender Systems

    Xiang Wang, Shigang Quan, Tingzhen Chang, Kang Yang, Sitong Chen, Yabo Fan, Xingxing Wang, Zhaodian He

    cs.AI

    Deploying high-dimensional multimodal features in industrial recommender systems incurs substantial storage and latency overhead. Hard quantization is compact but introduces boundary distortion, whereas dense soft quantization couples representation quality to the limited storage budget. We propose Sparse Activation-based Residual Soft Quantization (SA-RSQ), which uses Top-K sparse routing and softmax weights to store compact (Index,...

    arxiv.org/abs/2608.22979 · PDF

  37. 37

    Budget-Constrained Embodied Perception: Four Resource Walls and a Pre-Registered Evaluation of Access-Structured Perception on Open Models at less than 31B

    Defu Lin, Wenhui Chen, Ziyao Lin, Jianlin Chen, Peiji Long, Chi Man Vong

    cs.AI

    Embodied multimodal agents must answer from growing observation streams under a fixed per-decision token budget. We formalize this constraint through four resource walls: a perceptual Shannon wall for bounded state, a horizon wall for query-independent frame selection, a round wall for non-adaptive retrieval, and a conditional composition wall for fixed-depth inference. We introduce ASP, a training-free wrapper for frozen multimodal models...

    arxiv.org/abs/2608.22975 · PDF

  38. 38

    Toward Effective and Reliable LLM Agents via Dynamic Ontology

    Xiaohui Zhang, Zequn Sun, Chengyuan Yang, Yuanning Cui, Lingbing Guo, Wei Hu

    cs.AI

    Large language model (LLM) agents rely heavily on knowledge encoded in model parameters or presented as unstructured context. In domain-specific tasks, this leaves important semantic connections implicit. This often results in incomplete evidence use and brittle multi-step decisions. Ontologies offer a way to externalize domain concepts and relations as machine-interpretable structures, but constructing task-usable ontologies traditionally...

    arxiv.org/abs/2608.22974 · PDF

  39. 39

    ParallelWorld: Test-Time Scaling for Embodied Reasoning

    Min Chen, Shengjun Zhang, Yuxin Li, Zhang Zhang, Xin Fei, Chong Xia, Yueqi Duan

    cs.AI · cs.CV

    Embodied Reasoning constitutes a fundamental capability of embodied intelligence, serving as the basis for autonomous perception, reasoning, and interaction within physical environments. Recent studies have shifted the paradigm of embodied reasoning from static perception toward dynamic exploration, where agents acquire task-relevant information through interactions with the environment. However, existing active reasoning approaches generally...

    arxiv.org/abs/2608.22971 · PDF

  40. 40

    Buried in Textual Debt: Context Pruning with Visual Evidence Preservation for MLLM Agents

    Yuchen Huang, Sijia Li, Jun Zhang, Yi R. Fung

    cs.AI · cs.CL

    Multimodal Large Language Models (MLLMs) are increasingly deployed as multi-step agents, where explicit reasoning supports task decomposition and tool coordination but also accumulates self-generated text. Over long trajectories, this text can dominate the context and suppress visual evidence, creating textual debt. We observe that reasoning becomes redundant once task-relevant visual evidence is grounded, while stale hypotheses can misguide...

    arxiv.org/abs/2608.22963 · PDF

  41. 41

    What Process Evaluation of Coding Agents Actually Measures: Action, Task, and Step Are Three Different Levels

    Jiawei He, Mengyu Shi, Jie jia, Xikai Yang, Dong Sun

    cs.AI

    Coding agents are increasingly evaluated not only by whether they solve a task, but also by how they execute it. However, existing process-level evaluations often treat action prediction, task uncertainty, and step attribution as if they were the same problem, which makes it unclear what such evaluations actually measure. In this paper, we introduce a measurement framework for process evaluation in coding agents and instantiate step-level...

    arxiv.org/abs/2608.22960 · PDF

  42. 42

    Concepts for Securing Agentic AI Coding and the Terok Environment

    Jiří Vyskočil, Franz Pöschel, Andreas Knüpfer

    cs.AI · cs.SE

    Agentic AI is a fascinating new tool for software development. It is a huge step forward compared to "conventional" AI assisted coding, which in turn was a considerable breakthrough earlier. AI support through LLMs is a young and very fast-moving field. The "conventional" (non-agentic) flavor became useful and productive in early 2025 (around 18 months ago) and the agentic flavor followed in fall 2025 (approximately 9 months ago). Besides all...

    arxiv.org/abs/2608.22930 · PDF

  43. 43

    Beyond Observed Auxiliary Relations: Environment-Conditioned Modeling for Multi-Behavior Recommendation

    Seunghan Lee, Hyunsik Yoo, Jian Kang, Susik Yoon, SeongKu Kang

    cs.AI · cs.LG

    Multi-behavior recommendation (MBR) leverages auxiliary behavioral signals, such as clicks and add-to-cart, to enhance target behavior prediction like purchases. While recent graph neural network-based approaches have achieved strong performance by systematically propagating auxiliary behavior signals, they still suffer from two fundamental challenges inherent to auxiliary behaviors: (1) missing auxiliary signals, which hinder generalization...

    arxiv.org/abs/2608.22920 · PDF

  44. 44

    CDEG: Learning Decision-Critical Evidence for Long-Horizon Diagnostic Agents

    Xiwei Dai, Zijie Meng, Zhiting Fan, Yixuan Tang, Ziru Niu, Zuozhu Liu

    cs.AI

    Unlike static medical question answering, long-horizon diagnosis captures the sequential nature of clinical practice: evidence is progressively acquired, integrated, and evaluated over multiple rounds of interaction before reaching a final diagnosis. However, existing doctor agents often fail when critical evidence is either not acquired or not adequately incorporated into diagnostic reasoning. Recent agentic approaches attempt to address...

    arxiv.org/abs/2608.22899 · PDF

  45. 45

    Proxy reliance in large language model decisions is uncalibrated to predictive evidence

    Zengqing Wu, Chuan Xiao

    cs.AI · cs.CL · cs.CY

    Large language models (LLMs) are entering decisions in triage and lending, where task-relevant inference must be distinguished from impermissible proxy use. Current audits ask whether decisions change when demographics change. But attributes correlated with a protected group carry predictive value, so a changed decision can be discrimination or sound inference. We measure causal proxy effects in four LLMs on a clinical-ranking task with known...

    arxiv.org/abs/2608.22887 · PDF

  46. 46

    Your AI, On a Dial: Controlling Investment Bias in LLMs with a Single Neuron

    Sahong Park, Suhwan Park, Hoyoung Lee, Gakyung Kwon, Wonbin Ahn, Jaewon Choi, Alejandro Lopez-Lira, Yoon Kim,...

    cs.AI · cs.CL · q-fin.GN

    Large language models (LLMs) are increasingly used in investment decision-making, yet prior work shows that they exhibit systematic, model-specific investment preferences. We study whether a model's overall investment stance can be calibrated to a specified direction and strength. We introduce an investment-bias dial, an inference-time intervention on a single neuron that continuously adjusts a model-level decision prior---its overall...

    arxiv.org/abs/2608.22852 · PDF

  47. 47

    GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis

    Long Zhang, Yuhan Chen, Chaoran Zhang, Wanxia Cao, Kun Huang, Pengzhi Gao, Wei Liu, Jian Luan, Chenliang Li, Lixin Zou

    cs.AI

    Vision-Language Models (VLMs) based GUI agents stand to benefit significantly from online reinforcement learning (RL). However, their training is bottlenecked by two fundamental issues: current data synthesis methods for GUI Agents rely on specific environments and struggle to generate diverse data, while existing evaluators either suffer from limited scalability or provide inaccurate and unreliable reward signals. To overcome these...

    arxiv.org/abs/2608.22847 · PDF

  48. 48

    FinixDoc: Rethinking Financial Document Parsing Beyond Saturated Benchmarks

    Hang Wang, Jin Zhang, Guoliang Xu, Pengyue Lu, Yao Li, Zijiao Zhang, Tianyu Huang, Weiqi Xiong, Yulong Wang, Chuqiao...

    cs.AI

    Financial document parsing requires accuracy, structural consistency, and verifiability that current benchmarks often fail to reflect. We present FinixDoc, an end-to-end agentic parsing system for real-world financial documents, with FinixDoc-VL, a 4B-scale vision-language model built on Qwen3-VL-4B, as its core parser. To characterize the gap between benchmark and deployment performance, we introduce a Document Parsing Capability Matrix...

    arxiv.org/abs/2608.22842 · PDF

  49. 49

    Let the Bullets Fly: Multimodal Fake News Detection with Temporal-Aligned Generative Danmaku

    Xiansheng Luo, Chaowei Zhang, Zewei Zhang, Yi Zhu, Jipeng Qiang

    cs.AI

    The social interactions among crowds via \textit{Danmaku} (a.k.a., bullet comments) on modern multimedia platforms can facilitate both viewpoint conflicts and consensus, providing fine-grained discriminative social signals that can benefit fake news detection. However, the inherent accumulation latency of \textit{Danmaku} in real-world scenarios violates the real-time necessity of fake news detection, making the studies of...

    arxiv.org/abs/2608.22832 · PDF

  50. 50

    Beyond the Harness: End-to-End Optimization of Context Artifacts for Enterprise Text-to-SQL

    Kate Gwimm, Carson Eisenach

    cs.AI · cs.LG

    Deploying LLMs for enterprise Text-to-SQL is bottlenecked less by the model than by what context reaches it: business logic spans thousands of tables, and no model can ingest a full catalog at once. We argue that the most effective place to intervene is therefore the \emph{knowledge-base context} the model consumes, and that this context should be \emph{constructed} from historical usage rather than tuned for as a fixed input. Using a...

    arxiv.org/abs/2608.22830 · PDF

  51. 51

    Performance of a domain-specific large language model in answering patient questions in psychiatry

    Alexander J. Hish, Arjun Nagendran, Scott N. Compton

    cs.AI

    Background This study was designed to evaluate whether a domain-specific large language model (LLM) trained exclusively on patient education resources can answer questions about psychiatric medications, in a manner superior to LLM chatbots. We developed an LLM ("MIND") fine-tuned for clinical fidelity, trained on patient education resources from authoritative medical organizations. Methods We compared the responses of MIND, ChatGPT, and...

    arxiv.org/abs/2608.22797 · PDF

  52. 52

    TailSieve: Partial-Rollout-Guided Tail Routing for LLM Rollouts

    Tianqi Xu, Lu Lv, Haoyang Huang, Wenjie Huang, Zhanming Shen, Yuhao Shen, Baolin Zhang, Xinyi Hu, Shuang Ge, Jun...

    cs.AI · cs.LG

    Large-scale rollouts have become a core component of modern LLM systems, spanning reinforcement learning (RL) post-training, on-policy distillation (OPD), and sampling-heavy evaluation pipelines. Unlike online serving, which is typically optimized for request-level latency and throughput, a small number of long-tail generations can dominate the end-to-end makespan of an entire rollout step. In practice, rollout requests are often routed...

    arxiv.org/abs/2608.22788 · PDF

  53. 53

    The Retriever Should Remember: Experience-Amortized Reranking for Long-Term Agent Memory

    Qi Feng, Chris Ding, Jicong Fan

    cs.AI

    Long-term language-model agents accumulate memories across interactions, but their retrievers typically do not accumulate retrieval experience. Semantic retrieval is efficient, but embedding similarity does not always reflect whether a memory contains evidence relevant to the current query. Large language model (LLM) rerankers provide stronger query-conditioned relevance scores, yet stateless reranking repeatedly scores a large candidate pool...

    arxiv.org/abs/2608.22767 · PDF

  54. 54

    Compositional Chain-of-Relations for Faithful Knowledge Graph Question Answering with Large Language Models

    Chenhui Liu, Jianpeng Zhou, Jiahai Wang

    cs.AI

    Knowledge graph question answering (KGQA) is a key task for evaluating KG-augmented Large Language Models (LLMs), and complex KGQA that requires multi-hop reasoning is especially challenging. Solving a complex query involves two coupled phases: candidate retrieval, which locates answer candidates over the KG, and constraint handling, which filters these candidates against the query constraints. Faithful reasoning requires grounding both...

    arxiv.org/abs/2608.22762 · PDF

  55. 55

    The Compaction Cliff in Long-Running AI Agent Memory

    Saber Zerhoudi, Jelena Mitrovic, Michael Granitzer

    cs.AI · cs.IR

    A safety rule and an episodic log compete for the same tokens in an AI agent's context. When the budget overflows, both are summarized at the same rate; only the rule needs exact wording to remain enforceable. On 20 production agent configurations, Claude Code's /compact prompt on Sonnet 4.6 preserves 53\% of safety rules after one compaction round and 10\% after five. We name this the Compaction Cliff. We address it with Knowledge Triage, a...

    arxiv.org/abs/2608.22752 · PDF

  56. 56

    LLM-Based Selection of Incongruent Verbal and Nonverbal Behavior for Virtual Humans

    Parisa Ghanad Torshizi, Stacy Marsella

    cs.AI · cs.HC · cs.RO

    Nonverbal behavior generation systems for virtual agents often take an utterance as input and generate nonverbal behaviors that emphasize or illustrate the content of the verbal channel. However, human nonverbal behavior is shaped by more than the content of the speech. It is also influenced by speaker roles, interpersonal relationships, social context, and the cognitive and emotional states of the interactants. As a result, the nonverbal...

    arxiv.org/abs/2608.22731 · 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.