cs.AI（2026-06-12）

📊 共 93 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (60 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (19) 支柱一：机器人控制 (Robot Control) (7 🔗1) 支柱八：物理动画 (Physics-based Animation) (2) 支柱六：视频提取与匹配 (Video Extraction) (2 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (2 🔗1) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (60 篇)

#题目一句话要点标签🔗⭐
1 PersonaDrive: Human-Style Retrieval-Augmented VLA Agents for Closed-Loop Driving Simulation 提出PersonaDrive以解决闭环驾驶模拟中风格多样性不足的问题 vision-language-action VLA
2 Multi-Modal Agents for Power Distribution Defect Detection: An Evaluation of Foundation Models 提出多模态智能体框架以解决电力分配缺陷检测问题 foundation model multimodal
3 CMI-RewardBench: Evaluating Music Reward Models with Compositional Multimodal Instruction 提出CMI-RewardBench以评估音乐奖励模型的多模态指令 multimodal
4 Prefill Awareness in Large Language Models 提出预填意识以解决语言模型安全性问题 large language model
5 GeoNatureAgent Benchmark: Benchmarking LLM Agents for Environmental Geospatial Analysis Across Frontier and Open-Weight Foundation Models 提出GeoNatureAgent基准以评估环境地理空间分析的LLM代理 foundation model
6 PRISMR: Overcoming Parse Collapse in Multimodal Listwise Ranking via Parameterized Representation Internalization 提出PRISMR以解决多模态列表排名中的解析崩溃问题 multimodal
7 ARMOR-MAD: Adaptive Routing for Heterogeneous Multi-Agent Debate in Large Language Model Reasoning 提出ARMOR-MAD以解决多代理辩论中的计算浪费问题 large language model
8 Why Sampling Is Not Choosing: Intentionality, Agency, and Moral Responsibility in Large Language Models 探讨大型语言模型的意图性与道德责任问题 large language model
9 Automated reproducibility assessments in the social and behavioral sciences using large language models 利用大型语言模型自动化社会与行为科学的可重复性评估 large language model
10 Occupational Prompting Reveals Cultural Bias in Large Language Models 通过职业提示揭示大型语言模型中的文化偏见 large language model
11 Towards Data-free and Training-free Compression for Speech Foundation Models Using Parameter Clustering 提出无数据和无训练的语音基础模型压缩方法以解决模型体积问题 foundation model
12 MLUBench: A Benchmark for Lifelong Unlearning Evaluation in MLLMs 提出MLUBench以解决多模态大语言模型的终身遗忘问题 large language model multimodal
13 TerraBench: Can Agents Reason Over Heterogeneous Earth-System Data? 提出TerraBench以解决异构地球系统数据推理问题 large language model foundation model
14 Hallucination in Medical Imaging AI: A Cross-Modality Analytical Framework for Taxonomy, Detection, and Mitigation under Regulatory Constraints 提出跨模态分析框架以解决医学影像AI中的幻觉问题 foundation model chain-of-thought
15 Creating and Evaluating K-12 GenAI Assessment Graders Through Context Engineering 提出基于生成AI的K-12评估评分系统以提升教育评估效率 large language model foundation model
16 Adaptive Turn-Taking for Real-time Multi-Party Voice Agents 提出ModeratorLM以解决多方语音代理的轮流发言问题 large language model chain-of-thought
17 A Lightweight Multi-Agent Framework for Automated Concrete Barrier Design 提出轻量级多智能体框架以实现自动化混凝土障碍物设计 large language model
18 An Explainable AI Assistant for Introductory Programming Education: Improving Feedback Reliability with Instructor-AI Collaboration 提出可解释AI助手以提升编程教育中的反馈可靠性 large language model
19 ToolSense: A Diagnostic Framework for Auditing Parametric Tool Knowledge in LLMs 提出ToolSense框架以审计LLM中的工具知识检索问题 large language model
20 Zero-source LLM Hallucination Detection with Human-like Criteria Probing 提出人类标准探测方法以解决零源LLM幻觉检测问题 large language model
21 "Did you lie?" Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms 提出新型谎言检测器以解决模型信念验证问题 chain-of-thought
22 (Human) Attention Is (Still) All You Need: Human oversight makes AI-assisted social science reliable 提出人机协作经济研究框架以提升AI辅助社会科学的可靠性 large language model
23 DailyReport: An Open-ended Benchmark for Evaluating Search Agents on Daily Search Tasks 提出DailyReport以评估搜索代理在日常搜索任务中的表现 large language model
24 HarnessBridge: Learnable Bidirectional Controller for LLM Agent Harness 提出HarnessBridge以解决长任务中代理与环境交互的控制问题 large language model
25 Structured Testbench Generation for LLM-Driven HDL Design and Verification-Oriented Data Curation 提出结构化测试平台生成框架以解决LLM驱动的RTL设计验证问题 large language model
26 The Illusion of Multi-Agent Advantage 提出自动生成多智能体系统以解决单智能体系统的评估不足问题 chain-of-thought
27 Under What Conditions Can a Machine Become Genuinely Creative? 提出机器创造力的十项要求以解决人机协作问题 foundation model
28 Brick: Spatial Capability Routing for the Mixture-of-Models (MoM) Paradigm 提出Brick以解决大规模模型路由中的查询难度评估问题 multimodal
29 MiniMax Sparse Attention 提出MiniMax稀疏注意力以解决超长上下文处理问题 multimodal
30 A Three-Layer Framework for AI in Scientific Discovery 提出三层框架以推动科学发现中的AI应用 large language model
31 Agents-K1: Towards Agent-native Knowledge Orchestration 提出Agents-K1以解决科学知识编排不足的问题 multimodal
32 AI SciBrief as a Gateway to Research: A Framework for Onboarding Students into New Research Areas 提出AI SciBrief框架以解决学生研究入门障碍 large language model
33 Divination by Prompt: LLM-Mediated Xuanxue on Chinese Social Media 研究LLM介导的玄学实践以应对现代社会不确定性 large language model
34 Position: Generative Engine Optimization Creates Underexamined Risks, Governance Must Target Concentration, Disclosure, and Academic Blind Spots 提出生成引擎优化以应对信息获取中的潜在风险 large language model
35 HybridCodeAuthorship: A Benchmark Dataset for Line-Level Code Authorship Detection 提出HybridCodeAuthorship以解决代码作者检测的挑战 large language model
36 CAPED: Context-Aware Privacy Exposure Defense for Mobile GUI Agents 提出CAPED以解决移动GUI代理的隐私暴露问题 multimodal
37 PI-Hunter: Automated Red-Teaming for Exposing and Localizing Prompt Injections 提出PI-Hunter以解决大语言模型的潜在提示注入攻击问题 large language model
38 Beyond Problem Solving: UOJ-Bench for Evaluating Code Generation, Hacking, and Repair in Competitive Programming 提出UOJ-Bench以评估代码生成、黑客攻击和修复能力 large language model
39 Democracy in the Era of Artificial Intelligence 探讨人工智能如何提升民主制度的韧性与参与度 large language model
40 Functional Cache Grafting: Robust and Rapid Code-Policy Synthesis for Embodied Agents 提出功能缓存嫁接以解决代码策略生成的延迟与鲁棒性问题 large language model
41 Who Pays the Price? Stakeholder-Centric Prompt Injection Benchmarking for Real-world Web Agents 提出以利益相关者为中心的基准以解决提示注入攻击问题 large language model
42 Mod-Guide: An LLM-based Content Moderation Feedback System to Address Insensitive Speech toward Indigenous Ethnic and Religious Minority Communities 提出Mod-Guide以解决对少数民族不敏感言论的内容审核问题 large language model
43 AgentRivet: an automated system for producing Rivet routines from journal publications 提出AgentRivet以自动生成Rivet例程解决分析覆盖不足问题 large language model
44 DecompSR: A dataset for decomposed analyses of compositional multihop spatial reasoning 提出DecompSR数据集以分析组合多跳空间推理能力 large language model
45 DSAEval: Evaluating Data Science Agents on a Wide Range of Real-World Data Science Problems 提出DSAEval以评估数据科学代理在真实问题上的表现 multimodal
46 Epistemic Constitutionalism Or: how to avoid coherence bias 提出AI的知识宪法以解决一致性偏见问题 large language model
47 LLMs as ASP Programmers: Self-Correction Enables Task-Agnostic Nonmonotonic Reasoning 提出LLM+ASP框架以解决非单调推理问题 large language model
48 PhononBench:A Large-Scale Phonon-Based Benchmark for Dynamical Stability in Crystal Generation 提出PhononBench以解决晶体生成中的动态稳定性评估问题 large language model
49 HalluJudge: A Reference-Free Hallucination Detection for Context Misalignment in Code Review Automation 提出HalluJudge以解决代码审查自动化中的幻觉检测问题 large language model
50 DCD: Domain-Oriented Design for Controlled Retrieval-Augmented Generation 提出DCD以解决异构语料和多步查询的RAG质量下降问题 large language model
51 From Prompts to Responses: Dual-Sided Data Leakage and Defense in Split Large Language Models 提出双面数据泄露防御机制以解决分割大语言模型隐私问题 large language model ✅
52 A Two-Stage Statistical Framework for Evaluating Associative Interference in Large Language Models 提出两阶段统计框架以评估大语言模型中的关联干扰 large language model
53 Abstracting Cross-Domain Action Sequences into Interpretable Workflows 提出WorkflowView框架以解决跨域行为序列抽象问题 large language model
54 When the Tool Decides: LLM Agents Defer Blindly to Graph Neural Network Tools, and Stronger Backbones Defer More 研究表明LLM代理在使用GNN工具时缺乏判断力 large language model
55 GitOfThoughts: Version-Controlled Reasoning and Agent Memory You Can Replay, Diff, and Merge 提出GitOfThoughts以解决推理过程缺乏版本控制的问题 large language model
56 AudioDER: A Deduplication-Enhanced Reasoning Dataset for Post-Training Large Audio-Language Models 提出AudioDER以解决音频推理数据集冗余问题 chain-of-thought
57 Dense Coordinate-List Fine-Tuning Induces a Controllable Interference Surface in Vision-Language Models 提出密集坐标列表微调以控制视觉语言模型的干扰面 visual grounding
58 MeEvo: Metacognitive Evolution Combined with Natural Evolution for Automatic Heuristic Design 提出MeEvo以解决现有自动启发式设计方法的不足 large language model
59 VeriGeo: Controllable Geometry Question Generation with Numerical and Analytical Verification 提出VeriGeo以解决几何问题生成的可靠性与可控性问题 multimodal
60 FactoryLLM: A Safe and Open-Source AI Playground for Evaluating LLMs in Smart Factories 提出FactoryLLM以解决智能工厂中的故障诊断问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (19 篇)

#题目一句话要点标签🔗⭐
61 Reasoning for Mobile User Experience with Multimodal LLMs: Task, Benchmark, and Approach 提出UXBench基准以提升多模态大语言模型的用户体验评估能力 reinforcement learning large language model foundation model
62 A Tutorial on World Models and Physical AI 提出世界模型以解决物理AI中的智能决策问题 world model world models foundation model
63 Contextual Invertible World Models: A Neuro-Symbolic Agentic Framework for Colorectal Cancer Drug Response 提出上下文可逆世界模型以解决精准肿瘤学中的药物反应问题 world model world models large language model
64 IterCAD: An Iterative Multimodal Agent for Visually-Grounded CAD Generation and Editing 提出IterCAD以解决CAD生成与编辑中的迭代交互问题 reinforcement learning multimodal
65 Frozen Multimodal Embeddings for AI-Assisted Interview Assessment of Personality and Cognitive Ability 提出冻结多模态嵌入以解决心理特征与认知能力评估问题 representation learning multimodal
66 Topical Phase Transitions in Artificial Intelligence Research: Large-Scale Evidence and an Early-Warning Signature for Emerging Topics 提出主题相变模型以揭示人工智能研究的动态变化 reinforcement learning world model world models
67 GeoDial: A Multimodal Conversational Tutoring Dataset for Geometry Problem-Solving with Visual Tutor Turns 提出GeoDial数据集以解决几何问题求解中的视觉教学不足问题 teacher-student multimodal
68 From Digital to Physical: Digital Agents as Autonomous Coaches for Physical Intelligence 提出EmboCoach-Bench以解决机器人自主工程问题 diffusion policy reward design reward shaping
69 ReSum: Synergizing LLM Reasoning and Summarization with Reinforcement Learning 提出ReSum框架以优化大语言模型的推理与摘要过程 reinforcement learning large language model
70 Multi-Agent Reinforcement Learning from Delayed Marketplace Feedback for Objective-Weight Adaptation in Three-Sided Dispatch 提出一种多智能体强化学习方法以适应三方调度中的延迟市场反馈 reinforcement learning policy learning
71 Reasoning as Pattern Matching: Shared Mechanisms in Human and LLM Everyday Reasoning 提出模式匹配推理机制以解决人类与LLM推理一致性问题 world model world models large language model
72 Mental-R1: Aligning LLM Reasoning for Mental Health Assessment 提出CRPO框架以提升心理健康评估的推理能力 reinforcement learning large language model
73 EurekAgent: Agent Environment Engineering is All You Need For Autonomous Scientific Discovery 提出EurekAgent以解决自主科学发现中的环境设计问题 Eureka
74 Counterfactual Credit Policy Optimization for Multi-Agent Collaboration 提出反事实信用政策优化方法以解决多智能体协作中的信用分配问题 reinforcement learning large language model
75 Evoflux: Inference-Time Evolution of Executable Tool Workflows for Compact Agents 提出Evoflux以解决小型工具代理的执行效率问题 DPO distillation
76 The Theory of Mind Utility: Formal Specification of a Mentalizing Mechanism 提出理论心智效用模型以解决心智推理问题 world model world models
77 From Chatbot to Digital Colleague: The Paradigm Shift Toward Persistent Autonomous AI 提出从聊天机器人到数字同事的转变以实现自主AI的持续性工作 reinforcement learning large language model chain-of-thought
78 Causal Object-Centric Models for Planning with Monte Carlo Tree Search 提出COMET以提升基于蒙特卡洛树搜索的规划效率 reinforcement learning world model world models
79 Learning Coordinated Preference for Multi-Objective Multi-Agent Reinforcement Learning 提出PCMA以解决多目标多智能体强化学习中的决策协调问题 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (7 篇)

#题目一句话要点标签🔗⭐
80 From Verdict to Process: Agentic Reinforcement Learning for Multi-Stage Fact Verification 提出ProFact框架以优化多阶段事实验证流程 trajectory optimization reinforcement learning large language model
81 ComAct: Reframing Professional Software Manipulation via COM-as-Action Paradigm 提出COM-as-Action范式以解决专业软件操作的挑战 manipulation visual grounding
82 Physics-Guided Spatiotemporal Learning for Coastal Wave Peak Period Estimation from Video 提出物理引导的时空学习框架以估计海岸波峰周期 sim-to-real spatiotemporal
83 ERTS: Adversarial Robustness Testing of Ethical AI via Semantic Perturbation in a Bounded Consequence Space 提出ERTS框架以解决伦理AI的对抗鲁棒性测试问题 manipulation
84 CSPO: Constraint-Sensitive Policy Optimization for Safe Reinforcement Learning 提出约束敏感策略优化以解决安全强化学习问题 locomotion reinforcement learning
85 The Perceived Fragility of Explanations in Audio Models: Manipulation of Attribution with Unchanged Predictions 提出心理声学框架以解决音频深伪检测中解释脆弱性问题 manipulation ✅
86 Thinking Outside the [Chat]Box: Bridging Computer Science and Industrial Design for Cognitive-Inclusive Generative AI 提出双层支架框架以解决认知障碍者的生成式AI交互问题 manipulation multimodal

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗⭐
87 CloudCons: A Comprehensive End-to-End Benchmark for Cloud Resource Consolidation 提出CloudCons以解决云资源整合中的预测与决策问题 PULSE foundation model
88 TrajGenAgent: A Hierarchical LLM Agent for Human Mobility Trajectory Generation 提出TrajGenAgent以解决人类移动轨迹生成问题 spatiotemporal

🔬 支柱六：视频提取与匹配 (Video Extraction) (2 篇)

#题目一句话要点标签🔗⭐
89 Rethinking RAG in Long Videos: What to Retrieve and How to Use It? 提出V-RAGBench和CARVE以解决长视频检索生成问题 egocentric
90 StreamMemBench: Streaming Evaluation of Agent Memory for Future-Oriented Assistance 提出StreamMemBench以评估代理记忆在未来辅助中的有效性 egocentric ✅

🔬 支柱三：空间感知与语义 (Perception & Semantics) (2 篇)

#题目一句话要点标签🔗⭐
91 Generativism: Toward a Learning Theory for the Age of Generative Artificial Intelligence 提出Generativism以应对生成性人工智能时代的学习理论挑战 affordance
92 AFFORDANCE20Q: Evaluating Affordance Reasoning from Physical Properties 提出Affordance20Q以解决现有物体可供性推理评估不足的问题 affordance large language model ✅

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
93 The Emergence of Autonomous Penetration Capabilities in Large Language Model-Powered AI Systems 提出新的自主渗透评估框架以解决现有方法不足问题 penetration large language model

⬅️ 返回 cs.AI 首页 · 🏠 返回主页