cs.AI（2025-08-21）

📊 共 32 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (22 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (8 🔗1) 支柱一：机器人控制 (Robot Control) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (22 篇)

#题目一句话要点标签🔗⭐
1 GRAFT: GRaPH and Table Reasoning for Textual Alignment -- A Benchmark for Structured Instruction Following and Visual Reasoning GRAFT：提出图表推理基准，用于评估LLM在视觉文本对齐和结构化指令跟随上的能力。 multimodal instruction following
2 Foundation Models for Cross-Domain EEG Analysis Application: A Survey 脑电分析领域首个模态导向的预训练模型综述，填补了研究体系的空白。 foundation model multimodal
3 LLM4Sweat: A Trustworthy Large Language Model for Hyperhidrosis Support 提出LLM4Sweat，一个用于多汗症支持的可信大型语言模型框架 large language model foundation model
4 Flexible metadata harvesting for ecology using large language models 提出基于LLM的元数据收集器，解决生态数据整合中元数据异构问题 large language model
5 RETAIL: Towards Real-world Travel Planning for Large Language Models RETAIL：面向真实世界的大语言模型旅行规划数据集与主题引导多智能体框架 large language model
6 Invisible Filters: Cultural Bias in Hiring Evaluations Using Large Language Models 提出系统分析以解决大型语言模型招聘评估中的文化偏见问题 large language model
7 "Does the cafe entrance look accessible? Where is the door?" Towards Geospatial AI Agents for Visual Inquiries 提出Geo-Visual Agents，通过分析地理空间图像回答视觉空间查询。 multimodal
8 Language-Guided Tuning: Enhancing Numeric Optimization with Textual Feedback 提出语言引导调优（LGT），利用大语言模型和文本反馈提升数值优化效果。 large language model
9 Lean Meets Theoretical Computer Science: Scalable Synthesis of Theorem Proving Challenges in Formal-Informal Pairs 利用理论计算机科学，可扩展地合成形式化定理证明挑战 large language model
10 From Bits to Boardrooms: A Cutting-Edge Multi-Agent LLM Framework for Business Excellence BusiAgent：面向企业决策的多智能体LLM框架，提升战略规划与协作效率 large language model
11 Test-time Corpus Feedback: From Retrieval to RAG 综述性论文：探索测试时语料反馈在检索增强生成（RAG）中的应用，弥合IR与NLP的差距。 large language model
12 Multi-IaC-Eval: Benchmarking Cloud Infrastructure as Code Across Multiple Formats 提出Multi-IaC-Eval基准，评估LLM在多IaC格式上的代码生成与修改能力 large language model
13 ASIC-Agent: An Autonomous Multi-Agent System for ASIC Design with Benchmark Evaluation ASIC-Agent：用于ASIC设计的自主多智能体系统与基准评估 large language model
14 Noise, Adaptation, and Strategy: Assessing LLM Fidelity in Decision-Making 提出基于过程的评估框架，评估LLM在决策中模拟人类行为的保真度 large language model
15 HyperFlexis: Joint Design of Algorithms and Systems for Multi-SLO Serving and Fast Scaling HyperFlexis：面向多SLO服务和快速扩展的算法与系统联合设计 large language model
16 Cybernaut: Towards Reliable Web Automation Cybernaut：面向企业级应用的可靠Web自动化框架 large language model
17 Super-additive Cooperation in Language Model Agents 提出基于语言模型智能体的超加性合作博弈框架，提升多智能体协作能力 large language model ✅
18 IPIGuard: A Novel Tool Dependency Graph-Based Defense Against Indirect Prompt Injection in LLM Agents IPIGuard：一种基于工具依赖图的新型防御方法，用于抵御LLM Agent中的间接提示注入攻击 large language model
19 Coarse-to-Fine Grounded Memory for LLM Agent Planning 提出粗细粒度对齐记忆框架，提升LLM Agent在复杂规划任务中的适应性 large language model
20 M-$LLM^3$REC: A Motivation-Aware User-Item Interaction Framework for Enhancing Recommendation Accuracy with LLMs 提出M-$LLM^3$REC框架，利用大语言模型提取用户动机，提升推荐系统在冷启动场景下的准确性。 large language model
21 R-ConstraintBench: Evaluating LLMs on NP-Complete Scheduling 提出R-ConstraintBench以评估LLMs在NP完全调度问题上的表现 large language model
22 PuzzleClone: An SMT-Powered Framework for Synthesizing Verifiable Data PuzzleClone：一个基于SMT的框架，用于合成可验证的数据，提升LLM推理能力。 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
23 DeepThink3D: Enhancing Large Language Models with Programmatic Reasoning in Complex 3D Situated Reasoning Tasks DeepThink3D：通过程序化推理增强LLM在复杂3D场景理解中的能力 DPO direct preference optimization large language model
24 Think in Blocks: Adaptive Reasoning from Direct Response to Deep Reasoning 提出Think in Blocks框架，通过自适应推理块调整LLM的推理深度，提升效率。 reinforcement learning direct preference optimization large language model
25 SyGra: A Unified Graph-Based Framework for Scalable Generation, Quality Tagging, and Management of Synthetic Data SyGra：统一的图框架，用于合成数据的可扩展生成、质量标记和管理 DPO direct preference optimization large language model
26 Understanding Action Effects through Instrumental Empowerment in Multi-Agent Reinforcement Learning 提出基于工具性赋权的ICV方法，用于理解多智能体强化学习中的行为效应 reinforcement learning
27 NiceWebRL: a Python library for human subject experiments with reinforcement learning environments 提出NiceWebRL以支持人类参与的强化学习实验 reinforcement learning ✅
28 A Dynamical Systems Framework for Reinforcement Learning Safety and Robustness Verification 提出基于动力系统的强化学习安全与鲁棒性验证框架，保障安全关键系统应用。 reinforcement learning
29 An Empirical Study of Knowledge Distillation for Code Understanding Tasks 针对代码理解任务，系统性研究知识蒸馏的有效性与应用方法 distillation
30 Beyond Imaging: Vision Transformer Digital Twin Surrogates for 3D+T Biological Tissue Dynamics 提出VT-DTSN，利用Vision Transformer数字孪生替代网络预测3D+T生物组织动态。 predictive model distillation

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
31 Search-Based Credit Assignment for Offline Preference-Based Reinforcement Learning 提出基于搜索的信用分配方法，用于离线偏好强化学习 manipulation reinforcement learning offline reinforcement learning
32 The AI Model Risk Catalog: What Developers and Researchers Miss About Real-World AI Harms 构建AI模型风险目录，揭示开发者与研究者对AI危害认知的差异 manipulation

⬅️ 返回 cs.AI 首页 · 🏠 返回主页