cs.AI · 2026-05-30 · No. 12

Artificial Intelligence, 2026-05-30.

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

01 — The papers

48 entries

01

Physics Is All You Need? A Case Study in Physicist-Supervised AI Development of Scientific Software

Nhat-Minh Nguyen

cs.AI · astro-ph.CO · cs.HC · cs.SE

Are AI agents tools, co-authors, or researchers? We present a quantified case study ($N=1$): a physicist supervising an AI coding agent (Claude Code, Sonnet and Opus models) over 12 work days and 57 sessions to build CLAX-PT, a differentiable one-loop perturbation theory module in JAX. We documented and classified 15 supervision events by intervention level. The agent resolved ten autonomously by iterating against oracle tests. Two more by...

arxiv.org/abs/2605.30353 · PDF
02

SchGen: PCB Schematic Generation with Semantic-Grounded Code Representations

Qinpei Luo, Ruichun Ma, Xinyu Zhang, Lili Qiu

cs.AI · cs.CL · cs.LG

Printed circuit board (PCB) schematic design defines nearly all electronic hardware, but it remains manual and expertise-intensive. While generative AI has advanced digital and analog IC design, PCB schematic generation from natural-language intent is largely unexplored. This paper presents SchGen, the first large language model that generates editable PCB schematics from natural-language requests. The key challenge lies in the lack of an...

arxiv.org/abs/2605.30345 · PDF
03

Tiny but Trusted: Efficient Vision-Language Reasoning for Time-Series Anomaly Detection

Xiaona Zhou, Muntasir Wahed, Tianjiao Yu, Constantin Brif, Ismini Lourentzou

cs.AI

Recent advances in Vision-Language Models (VLMs) have achieved impressive performance across many tasks, yet prior studies report unsatisfactory performance when applying large language or multimodal models to finding abnormal patterns in sequential data. Public anomaly detection benchmarks typically provide interval annotations but not natural-language rationales, making it difficult to fine-tune VLMs to produce grounded, interpretable...

arxiv.org/abs/2605.30344 · PDF
04

Locally Coherent, Globally Incoherent: Bounding Compositional Incoherence in Multi-Component LLM Agents

Anany Kotawala

cs.AI · cs.CL

Multi-component LLM agents assemble probabilistic claims from components that each see only part of a joint problem; the composition can violate basic probability axioms even when every component is locally coherent. We formalise this locally coherent, globally incoherent failure via the compositional residual eps*, the L2 distance from the composed quote to the joint coherent polytope, computable at runtime from system output and the...

arxiv.org/abs/2605.30335 · PDF
05

Demystifying Data Organization for Enhanced LLM Training

Yalun Dai, Yangyu Huang, Tongshen Yang, Yonghan Wang, Xin Zhang, Wenshan Wu, Qihao Zhao, Hao Li, Yuanyuan Gao,...

cs.AI · cs.CL

Large Language Models (LLMs) have revolutionized various fields, yet their training efficiency is heavily reliant on effective data curation. While data selection has been widely studied, the strategic data organization for enhanced training remains an underexplored area, particularly since current LLMs are often trained for only one or a few epochs. This paper systematically explores the influence of data organization on LLM training by...

arxiv.org/abs/2605.30334 · PDF
06

MIRA: Mid-training Rubric Anchoring for Source-Aware Data Selection

Haowen Wang, Yaxin Du, Jian Yang, Jiajun Wu, Shukai Liu, Yuxuan Zhang, Pingjie Wang, Siheng Chen, Tuney Zheng, Ming...

cs.AI

Mid-training has become an important stage in modern LLM development, using large-scale curated mixtures to strengthen capabilities before final post-training. Its data selection problem is distinct: the data are optimized under a pretraining-style objective at near-pretraining scale, but are curated toward downstream capabilities and drawn from heterogeneous sources with different formats and training roles. As a result, effective selection...

arxiv.org/abs/2605.30288 · PDF
07

ProjectionBench: Evaluating Scientific Hypothesis Generation in LLMs Under Progressive Information Disclosure

A. J. Lew, Y. Cao, M. J. Buehler

cs.AI

Scientific discovery is an inherently creative and uncertain process, requiring reasoning beyond the recall of known knowledge. While many benchmarks have been proposed to evaluate large language model (LLM) performance on deep research tasks via multi-hop retrieval, their innovative reasoning abilities essential for true scientific discovery remain largely untested. We introduce a benchmark framework for evaluating model performance in...

arxiv.org/abs/2605.30284 · PDF
08

mcp-proto-okn: Natural-language access to open scientific knowledge graphs through the Model Context Protocol

Peter W. Rose, Benjamin M. Good, Amanda M. Saravia-Butler, Charlotte A. Nelson, James P. Balhoff, Yaphet Kebede,...

cs.AI · cs.ET

MCP Server Proto-OKN (mcp-proto-okn) is a Python-based Model Context Protocol server that enables AI assistants to discover, inspect, query and integrate scientific knowledge graphs through natural language. The server provides graph routing, schema inspection, SPARQL execution, ontology expansion, multi-graph querying, and transcript generation, lowering the barrier to cross-domain knowledge graph analysis for biomedical and scientific...

arxiv.org/abs/2605.30283 · PDF
09

When Should Models Change Their Minds? Contextual Belief Management in Large Language Models

Haoming Xu, Weihong Xu, Zongrui Li, Mengru Wang, Yunzhi Yao, Chiyu Wu, Jin Shang, Yu Gong, Shumin Deng

cs.AI · cs.CL · cs.LG

Long-horizon interactions require language models to manage accumulating information: when to update their state, when to preserve their state, and what to ignore. We study this challenge as \textbf{Contextual Belief Management (CBM)}: maintaining a predicted belief state aligned with formal evidence while isolating task-irrelevant noise. To make CBM measurable, we introduce BeliefTrack, a closed-world benchmark spanning Rule Discovery and...

arxiv.org/abs/2605.30219 · PDF
10

Persona Conditioning of Brand Recommendations in Retrieval-Augmented Commercial Chat: A Prominence-Stratified Cross-Provider Audit

Will Jack, Noah Lehman, Keller Maloney, Sarah Xu

cs.AI

The same prompt -- "best CRM software" -- reaches AI assistants from buyers in widely different contexts: a solo founder, an enterprise VP, a UK SMB owner. We audit how strongly that contextual variation reshapes which brands the model recommends. The audit samples 2,000 runs over a design space of 10 personas x 8 prompts x 3 model configurations x N=10 reps, with the two OpenAI cells at full 8-prompt coverage and the Anthropic sonnet-4.6 /...

arxiv.org/abs/2605.30207 · PDF
11

Double-Edged Sword or Sharp Tool? Designing and Evaluating Triadic LLM-Teacher Collaboration for K-12 Writing at Scale

Canran Wang, Yuwen Yang, Zhen Wang, Ming Ma, Ding Yu, Chentai Wang, Keman Huang, Xiaoyong Du

cs.AI

The double-edged sword of integrating Large Language Models (LLMs) requires an effective triadic collaboration mechanism among LLMs, teachers and students, especially for K-12 education. By developing a triadic collaboration system to support K-12 writing learning, a multidimensional evaluation framework grounded in Systemic Functional Linguistics and the suggestion trajectory tracing pipeline, this paper contributes a large-scale empirical...

arxiv.org/abs/2605.30200 · PDF
12

Modularizing Educational LLM-Agency for Fostering Responsible Learning Assistance

Julius Gabelmann, Felix Jahn, Kevin Baum, Sophie van Rossum, Emely Wuenscher, Timo P. Gros, Verena Wolf

cs.AI · cs.CY

The widespread adoption of AI chatbots in education will drastically change learning, making responsible deployment a critical concern. While large language models (LLMs) might have access to sources discussing insights from educational sciences, they are not particularly inclined to adhere to pedagogical concepts, risking negative effects on the learning process, such as a loss of transfer capabilities, critical thinking, or creativity. In...

arxiv.org/abs/2605.30187 · PDF
13

BioRefusalAudit: Auditing Biosecurity Refusal Depth Using General and Domain-Fine-Tuned Sparse Autoencoders

Caleb DeLeeuw

cs.AI · cs.CR · cs.LG

Biosecurity evaluations of language models typically ask whether models produce hazardous output. This paper asks a complementary question: when a model refuses, is that refusal structurally sound, or does it disappear under modest changes to prompt framing, formatting, or output length? Across five architectures, no model cleanly discriminated benign from hazard. Gemma 2 2B-IT never genuinely refused across 75 prompts, hedging on every...

arxiv.org/abs/2605.30162 · PDF
14

Meta-Cognitive Memory Policy Optimization for Long-Horizon LLM Agents

Ziyan Liu, Zhezheng Hao, Yeqiu Chen, Hong Wang, Jingren Hou, Ruiyi Ding, Yongkang Yang, Wence Ji, Wei Xia, Feng Liu

cs.AI

Memory-augmented LLM agents tackle complex long-horizon tasks by recursively summarizing interaction trajectories into compact memory. However, existing approaches typically train these memory policies using outcome-based reinforcement learning, failing to localize where intermediate memory quality degrades. As interactions unfold, ambiguous recursive summaries progressively discard task-relevant information and introduce semantic noise. This...

arxiv.org/abs/2605.30159 · PDF
15

Temporal Stability and Few-Shot Prompting in Math Task Assessment

Danielle S. Fox, Brenda L. Robles, Elizabeth DiPietro Brovey, Christian D. Schunn

cs.AI

As AI tools become increasingly integrated into educational contexts, questions arise about both their stability over time and their responsiveness to prompt engineering techniques. This longitudinal study focused on different AI tools' ability to use the Task Analysis Guide (TAG; Stein \& Smith, 1998) to classify the cognitive demand of mathematics tasks. In particular, it examined whether this classification ability changed with (1) model...

arxiv.org/abs/2605.30151 · PDF
16

Anchorless Diversification for Parallel LLM Ideation

Fares Nabil Ibrahim, Nafis Saami Azad, Raiyan Abdul Baten

cs.AI

LLMs are increasingly used to generate candidate-idea pools for creative tasks where broad exploration is valuable. Parallel inference can be attractive in this setting when it broadens the pool while retaining quality and cost efficiency. We study inference-time controls for candidate-pool diversification, asking whether anchorless methods can rival methods that depend on observed seed ideas. Across three creative task families, we compare...

arxiv.org/abs/2605.30150 · PDF
17

AgentSchool: An LLM-Powered Multi-Agent Simulation for Education

Yulei Ye, Wenhao Li, Zhong Wen, Yunshu Huang, Yichen Hu, Zifan Wei, Yige Wang, Xinyu Xie, Haoxuan Yang, Yanjun...

cs.AI · cs.MA

Despite the rapid deployment of LLMs into classrooms, validating educational AI remains uniquely intractable: interventions act on developing learners whose cognitive and social trajectories are irreversibly shaped, while real-world trials are slow, ethically constrained, and institutionally locked. LLM-based educational simulators have emerged as a potential remedy, but many still collapse learning into persona-conditioned role-play and,...

arxiv.org/abs/2605.30144 · PDF
18

Enhancing Multi-Agent Communication through Attention Steering with Context Relevance

Hongxiang Zhang, Yuan Tian, Tianyi Zhang

cs.AI

LLM-based multi-agent systems have demonstrated remarkable performance on complex tasks through collaborative reasoning. However, these systems tend to rapidly accumulate extremely long conversation histories during interaction. As conversations lengthen, relevant information is increasingly diluted by irrelevant context, leading to degraded performance. In this work, we present Agent-Radar, a training-free context management method that...

arxiv.org/abs/2605.30136 · PDF
19

VLA-Trace: Diagnosing Vision-Language-Action Models through Representation and Behavior Tracing

Haoyuan Shi, Xiancong Ren, Yingji Zhang, Qinfan Zhang, Jiayu Hu, Haozhe Shan, Han Dong, Jinpeng Lu, Yinda Chen, Yi...

cs.AI

Understanding how Vision-Language-Action (VLA) models transform multimodal knowledge into embodied control remains an open challenge. We present VLA-Trace, a progressive diagnostic framework that analyzes VLA models through a unified evidence chain from representation dynamics to causal control attribution and behavioral manifestation. It specifically combines cross-modal and checkpoint-drift centered kernel alignment (CKA) to trace...

arxiv.org/abs/2605.30117 · PDF
20

PokerSkill: LLMs Can Play Expert-Level Poker without Training or Solvers

Boning Li, Baoxiang Wang, Longbo Huang

cs.AI · cs.GT

Poker is a landmark challenge for artificial intelligence. The dominant approach relies on equilibrium solvers built on counterfactual regret minimization, requiring millions of core-hours of training. Large Language Models (LLMs) possess extensive poker knowledge but perform far below solver-based agents when asked to play directly. Traditional rule-based poker agents are interpretable and training-free, but their strategic ceiling remains...

arxiv.org/abs/2605.30094 · PDF
21

Selective QA over Conflicting Multi-Source Personal Memory: A Diagnostic Testbed and Method Comparison

Tiancheng Yang, Matthias Schonlau, Ilia Sucholutsky

cs.AI

Emerging personal AI agents are moving toward persistent, multi-source memory. This creates an evaluation problem: systems must decide how to use conflicting or incomplete evidence; they cannot just retrieve facts from one clean history. Existing benchmarks rarely show whether an error came from the evidence given to a method or from the method's conflict-resolution step. We study this as selective QA over conflicting multi-source personal...

arxiv.org/abs/2605.30087 · PDF
22

Conformal Certification of Reasoning Trace Prefixes

Matt Y. Cheung, Ashok Veeraraghavan, Hanjie Chen, Guha Balakrishnan

cs.AI · cs.CL · cs.LG · stat.ML

Language model reasoning traces are rarely all-or-nothing; they frequently contain valid intermediate steps before a critical error occurs. Existing uncertainty quantification methods typically certify final answers or entire responses, failing to provide statistical guarantees for the proportion of a sequential trace that can be safely retained. To address this, we introduce CROP (Conformal Reasoning Output Prefixes), a verifier-agnostic...

arxiv.org/abs/2605.30085 · PDF
23

Robust and Generalizable Safety Steering for Text-to-Image Diffusion Transformers

Zihao Xue, Yan Wang, Zhen Bi, Long Ma, Zhonglong Zheng, Zeyu Yang, Bingyu Zhu, Longtao Huang, Jie Xiao, Jungang Lou

cs.AI

Diffusion Transformers have become a powerful backbone for text-to-image generation, but their layered and cross-modal generation process makes safety control fundamentally different from prompt-level filtering or output-level detection. Harmful semantics may be weakly expressed in text representations, progressively bound to visual latents, and finally entangled with rendering dynamics. As a result, safety steering at a fixed layer can be...

arxiv.org/abs/2605.30049 · PDF
24

Learning to Choose: An Empowerment-Guided Multi-Agent System with semantic communication for Adaptive Method Selection

Geremy Loachamín-Suntaxi, Robert Lazar, Dimitrios G. Giovanis, Ioannis G. Kevrekidis, Eleni D. Koronaki

cs.AI

Automating scientific computing workflows requires more than generating executable code: autonomous systems must also select appropriate computational strategies, implement them faithfully, and ensure that the resulting outcomes remain causally attributable to the decisions that produced them. In multi-agent pipelines, this process is particularly fragile, as small inconsistencies between agent intentions and actions can lead to semantic...

arxiv.org/abs/2605.30042 · PDF
25

Domain-Specific Data Synthesis for LLMs via Minimal Sufficient Representation Learning

Tong Ye, Hang Yu, Tengfei Ma, Xuhong Zhang, Jianguo Li, Peng Di, Peiyu Liu, Jianwei Yin, Wenhai Wang

cs.AI

Large Language Models have demonstrated remarkable progress in general-purpose capabilities and can achieve strong performance in specific domains through fine-tuning on domain-specific data. However, acquiring high-quality data for target domains remains a significant challenge. Existing data synthesis approaches follow a deductive paradigm, heavily relying on explicit domain descriptions expressed in natural language and careful prompt...

arxiv.org/abs/2605.30039 · PDF
26

Teaching Values to Machines: Simulating Human-Like Behavior in LLMs

Asaf Yehudai, Naama Rozen, Ariel Gera

cs.AI · cs.CL

Large Language Models (LLMs) demonstrate a remarkable capacity to adopt different personas and roles; however, it remains unclear whether they can manifest behavior that adheres to a coherent, human-like value structure. In this work, we draw on established psychological value theory to induce human-like values in LLMs and assess their alignment with patterns observed in human studies. Using validated psychological questionnaires, we conduct...

arxiv.org/abs/2605.30036 · PDF
27

RAISE: RAG Design as an Architecture Search Problem

Zhen Chen, Yibing Liu, Weihao Xie, Yu Liang, Peilin Chen, Shiqi Wang

cs.AI

Retrieval-augmented generation (RAG) systems expose numerous design choices spanning query rewriting, chunking, retrieval depth, reranking, and context compression. In practice, these choices are often configured through heuristics, hindering systematic evaluation and reproducibility across settings. We argue that this challenge is best formulated as RAG architecture search. To support controlled and reproducible study of this problem, we...

arxiv.org/abs/2605.30029 · PDF
28

From GPS Points to Travel Patterns: Flexible and Semantic Trajectory Generation with LLMs

Silin Zhou, Chenhao Wang, Yuntao Wen, Shuo Shang, Lisi Chen, Panos Kalnis

cs.AI

Urban trajectories play a crucial role in modeling urban dynamics and supporting various smart city applications. However, privacy concerns restrict access to large-scale and high-quality trajectory datasets. Trajectory generation provides a promising alternative by synthesizing realistic data to mitigate privacy risks. However, existing methods fail to explicitly capture travel patterns and can only generate fixed-length trajectories under a...

arxiv.org/abs/2605.30014 · PDF
29

KairosAgent: Agentic Time Series Forecasting with Fused Semantic Reasoning

Kun Feng, Ziwei Shan, Yuchen Fang, Yiyang Tan, Sihan Lu, Shuqi Gu, Lintao Ma, Xingyu Lu, Kan Ren

cs.AI

Cross-domain multimodal time series forecasting is a challenging task, requiring models to integrate precise numerical comprehension, cross-domain semantic understanding, and effective multimodal fusion. Existing approaches either build Time Series Foundation Models (TSFMs) from scratch or leverage pretrained Large Language Models (LLMs). However, TSFMs often overlook semantic understanding and lack the ability to perform future-oriented...

arxiv.org/abs/2605.30002 · PDF
30

Cookie-Bench: Continuous On-screen Key Interaction Evaluation for Web Generation

Haoyue Yang, Zhangxiao Shen, Fan Ding, Hangting Lou, Yifeng Kou, Haoqing Yu, Jingyao Li, Zhengfan Wu, Siqi Bao, Jing...

cs.AI

Front-end web code has become a core product surface for every frontier LLM release, yet evaluating these interactive applications at development speed remains costly because human-judged leaderboards like Arena do not scale. Existing automated proxies typically lean on reference implementations, test suites, or rigid checklists, and tend to miss the reasoned synthesis a human reviewer performs over a live session. We articulate a new...

arxiv.org/abs/2605.30000 · PDF
31

Accelerating Constrained Decoding with Token Space Compression

Michael Sullivan, Alexander Koller

cs.AI

To guarantee that an LLM's outputs conform to a specified structure, context-free grammar (CFG) decoding engines force the selection of next tokens that produce strings that conform to a given CFG. While current CFG-constrained decoding engines are highly optimized, the inherent costs arising from the massive per-step search space -- i.e. the entire token vocabulary -- result in intractably high overhead for more complex CFGs: precisely the...

arxiv.org/abs/2605.29986 · PDF
32

Compass: Navigating Global Marine Lead Data Integration through Expert-Guided LLM Agent

Yiming Liu, Bin Lu, Meng Jin, Ziyuan Sang, Shuo Jiang, Lei Zhou, Xinbing Wang, Chenghu Zhou, Jing Zhang

cs.AI

Marine lead (Pb) and its isotopes are critical tracers for ocean circulation and anthropogenic pollution, yet in-situ observations remain costly and sparse. While vast historical records exist, they lie buried within the unstructured content of academic papers, creating "data silos" inaccessible to comprehensive analysis. Manual extraction is unscalable, while general-purpose Large Language Models (LLMs) lack the necessary domain-specific...

arxiv.org/abs/2605.29966 · PDF
33

Meta-Programming for Linear-time Temporal Answer Set Programming

Susana Hahn, Amade Nems, Javier Romero, Torsten Schaub

cs.AI

The development of temporal extensions of Answer Set Programming (ASP) has led to the emergence of non-monotonic linear-time (TEL), dynamic (DEL), and metric (MEL) temporal equilibrium logics. However, the inherent rigidity of highly optimized ASP systems often hinders the rapid exploration and implementation of alternative logical designs. In this work, we propose a flexible meta-programming framework that operationalizes the semantics of...

arxiv.org/abs/2605.29965 · PDF
34

Formalizing Mathematics at Scale

Ahmad Rammal, Niket Patel, Fabian Gloeckle, Amaury Hayat, Julia Kempe, Remi Munos, Charles Arnal, Vivien Cabannes

cs.AI

We present AutoformBot, a multi-agent system for building an Autoformalized Textbook Library At Scale (Atlas) in Lean 4. AutoformBot orchestrates thousands of LLM agents, equipped with formal verification tools, dependency-aware task scheduling, and collaborative version control, to translate informal textbook prose into machine-checked definitions and proofs. We apply our methods to a corpus of 26 open-access textbooks spanning analysis,...

arxiv.org/abs/2605.29955 · PDF
35

MuPHI: Learning Implicit Multimodal Harm Reasoning via Semantically Grounded Reward Optimization

Anisha Saha, Varsha Suresh, Teodora Kamova, Sophia Wiedmann, Timothy Hospedales, Vera Demberg

cs.AI · cs.CL · cs.LG · cs.MM

Understanding how harm emerges from interaction between otherwise benign image-text pairs requires intent-aware cross-modal reasoning beyond surface-level features. Existing vision-language models (VLMs) excel at literal reasoning over perceptual cues but often fail to derive harmful semantics that rely on implicit, context-dependent reasoning. To evaluate VLMs on compositional harm detection and reasoning, we introduce Multimodal Pragmatic...

arxiv.org/abs/2605.29951 · PDF
36

Make LLM Learn to Synthesize from Streaming Experiences through Feedback

Zhenlin Hu, Yan Wang, Zhen Bi, Zihao Xue, Bingyu Zhu, Longtao Huang, Xiongtao Zhang, Zeyu Yang, Zhixuan Chu, Jungang Lou

cs.AI

Large language models (LLMs) have been widely adopted for synthetic data generation, significantly reducing annotation costs. However, most existing studies treat synthesis as a set of isolated tasks and overlook a more fundamental question: whether a model can learn to synthesize by accumulating experience from past tasks and transferring it to future ones. In this work, we introduce StreamSynth, a new setting in which synthesis tasks arrive...

arxiv.org/abs/2605.29940 · PDF
37

It`s All About Speed: AI`s Impact on Workflow in Music Production

Finn McClellan, Fabio Morreale

cs.AI · eess.AS

In this paper, we present the results of an ethnographic study into the impact of AI and automated tools on music production workflow. Focusing specifically on professional participants who identified as recording engineers, mixers, and producers, we discuss their usage of common AI and automated software, as well as their sentiments on the proliferation of these tools. We discuss tensions that may be created between users and automated tools...

arxiv.org/abs/2605.29931 · PDF
38

Toward AI Systems That Understand Self and Others: A Multi-Phase Inference Framework for Human Cognitive Diversity and World-Model Alignment

Toru Takahashi

cs.AI · cs.CY · cs.HC

Mutual misunderstanding in contemporary society does not arise merely because people hold different opinions or values. Even under the same observations, different subjects may form different inferential targets, state representations, prediction errors, and update priorities. This paper proposes a multi-phase inference framework and defines its core internal mechanism as the Multi-Phase Inference Mechanism (MIM). MIM formalizes how...

arxiv.org/abs/2605.29930 · PDF
39

On the Geometry of Games and their Solvers

Yaqi Sun, Julian Ma, David Mguni

cs.AI · cs.MA

A central challenge in game theory and learning systems such as GANs is understanding which algorithms can efficiently compute equilibria across the heterogeneous landscape of games. Equilibrium computation is typically studied solver by solver and game class by game class, yielding strong local guarantees but a fragmented view of solver behaviour. Existing discrete taxonomies often provide an incomplete account of where algorithms succeed....

arxiv.org/abs/2605.29919 · PDF
40

Redundant or Necessary? A Benchmark for Detecting Redundant Steps in Agent Trajectories

Minyang Hu, Bo Yang, Zhinuo Zhou, Jiachen Liang, Guo Jiahao, Yiyang Yin, Xiongwei Han

cs.AI

LLM-based agents have demonstrated strong capabilities in solving complex tasks through multi-step reasoning and tool use. However, existing evaluation protocols primarily focus on task success, overlooking a critical aspect of agent behavior: execution efficiency. In practice, agent trajectories often contain redundant steps that consume substantial resources while contributing little to task completion. In this work, we propose and...

arxiv.org/abs/2605.29893 · PDF
41

Moment-KV: Momentum-Based Decode-Time KV Cache Compression for Long Generation

Soumyadeep Jana, Sagar Nishad, Sanasam Ranbir Singh

cs.AI

Key-Value (KV) cache remains a major bottleneck for deploying Large Language Models (LLMs) in long-generation tasks. Prior work often applies uniform compression across both prefill and decoding caches, but compressing the prefill cache degrades performance by corrupting critical context. While preserving the prefill cache is essential, decoding-phase compression remains underexplored, with existing methods relying on rigid recency windows or...

arxiv.org/abs/2605.29873 · PDF
42

OmniMatBench: A Human-Calibrated Multimodal Reasoning Benchmark Across 19 Materials Science Subfields

Wanhao Liu, Jiaqing Xie, Qian Tan, Weida Wang, Jue Wang, Ran Sun, Zhuo Yang, Wanli Ouyang, Lei Bai, Tianfan Fu, Lu...

cs.AI

As multimodal language models play an increasingly important role in scientific research, materials science offers a critical testbed due to its interdisciplinary, multimodal, and application-driven nature. However, existing materials benchmarks mainly focus on property prediction, knowledge QA, or characterization understanding, leaving the broader reasoning process from materials knowledge to application underexplored. To fill this gap, we...

arxiv.org/abs/2605.29833 · PDF
43

OptSkills: Learning Generalizable Optimization Skills from Problem Archetypes via Cluster-Based Distillation

Haochen Yang, Ke Zhao, Mengyuan Ma, Xingyu Lu, Xiangfeng Wang, Hong Qian

cs.AI · cs.LG

Leveraging Large Language Models (LLMs) to automatically formulate and solve optimization problems from natural language has emerged as an efficient paradigm for automated optimization. However, existing methods still exhibit limited generalization: they are sensitive to superficial narrative variations, reuse experience mainly at the case level, and struggle to adapt to shifted or emerging problem types. We propose OptSkills, an...

arxiv.org/abs/2605.29829 · PDF
44

Quantifying and Optimizing Simplicity via Polynomial Representations

Tianren Zhang, Xiangxin Li, Minghao Xiao, Guanyu Chen, Feng Chen

cs.AI

Deep networks often exhibit a preference for "simple" solutions, and such a simplicity bias is widely believed to play a key role in generalization. Yet a broadly applicable, quantitative measure of simplicity remains elusive. We introduce polynomial representations as a distribution-aware, low-dimensional surrogate for neural functions: we approximate a network's predictive behavior along data-dependent interpolation paths using orthogonal...

arxiv.org/abs/2605.29823 · PDF
45

Harnessing non-adversarial robustness in large language models

Qinghua Zhou, Ellina Aleshina, Andrey Lovyagin, Oleg Somov, Mikhail Seleznyov, Alexander Panchenko, Ivan Oseledets,...

cs.AI

The work presents an approach for addressing the challenge of robustness in Large Language Models (LLMs) to alterations and potential errors caused by semantically similar but textually different prompts. Recent works have shown that these kinds of prompt variations can significantly impact the performance of LLMs on tasks. The central question is: can LLMs' robustness to semantically-neutral prompt alterations be acquired without expensive...

arxiv.org/abs/2605.29816 · PDF
46

PRAIB: Peer Review AI Benchmark of Behaviour of LLM-Assisted Reviewing

Krzysztof Żurawicki, Julia Farganus, Arkadiusz Gaweł, Mateusz Bystroński, Tomasz Jan Kajdanowicz

cs.AI · cs.CL

The growing number of submitted papers has motivated the exploration of Large Language Models (LLMs) as a means to support and augment the peer review process, particularly in terms of improving its speed and scalability. Yet, it remains unknown whether LLMs engage with scientific manuscripts in the same manner as human reviewers, or whether they merely produce review-looking text. To address this, we introduce the Peer Review AI Benchmark...

arxiv.org/abs/2605.29815 · PDF
47

AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Security

Dongrui Liu, Yu Li, Zhonghao Yang, Peng Wang, Guanxu Chen, Yuejin Xie, Qinghua Mao, Wanying Qu, Yanxu Zhu, Tianyi...

cs.AI · cs.CL · cs.CR · cs.CV · cs.LG

Modern open-world agents such as OpenClaw exhibit powerful cross-environment execution capabilities yet introduce broad new safety risk sources. Meanwhile, advanced frontier AI models drastically lower attack barriers, rendering current agent alignment frameworks inadequate for real-world deployment. To tackle these emerging threats, we propose a lightweight and scalable agent safety alignment framework. Specifically, we update the agent...

arxiv.org/abs/2605.29801 · PDF
48

SAAS: Self-Aware Reinforcement Learning for Over-Search Mitigation in Agentic Search

Yunbo Tang, Chengyi Yang, Shiyu Liu, Zhishang Xiang, Zerui Chen, Qinggang Zhang, Jinsong Su

cs.AI · cs.CL · cs.LG

Agentic search enables LLMs to solve complex multi-hop questions through iterative reasoning and external search. Despite the effectiveness, these systems often suffer from a critical limitation in practice: agents fail to recognize their own knowledge boundaries, blindly triggering searches when internal knowledge suffices and failing to terminate search even when adequate evidence has been collected. The lack of self-awareness leads to...

arxiv.org/abs/2605.29796 · 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.