cs.AI（2026-02-09）

📊 共 33 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) (25 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗1) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (25 篇)

#题目一句话要点标签🔗⭐
1 Grounding Generative Planners in Verifiable Logic: A Hybrid Architecture for Trustworthy Embodied AI 提出VIRF框架，通过逻辑导师与LLM协作，实现具身AI可验证的安全规划。 embodied AI large language model
2 Gesturing Toward Abstraction: Multimodal Convention Formation in Collaborative Physical Tasks 研究人机协作中多模态沟通策略的演化，并提出基于概率模型的抽象概念形成方法。 multimodal
3 Scalable Delphi: Large Language Models for Structured Risk Estimation 提出Scalable Delphi，利用大语言模型实现可扩展的结构化风险评估 large language model
4 DeepQuali: Initial results of a study on the use of large language models for assessing the quality of user stories DeepQuali：利用大型语言模型评估用户故事质量的初步研究 large language model
5 Root Cause Analysis Method Based on Large Language Models with Residual Connection Structures 提出基于残差连接和大型语言模型的RC-LLM方法，用于微服务架构中的根因分析 large language model
6 Dynamics Within Latent Chain-of-Thought: An Empirical Study of Causal Structure 通过因果结构分析，揭示隐式思维链中的动态过程与决策机制 chain-of-thought
7 6G-Bench: An Open Benchmark for Semantic Communication and Network-Level Reasoning with Foundation Models in AI-Native 6G Networks 6G-Bench：面向AI原生6G网络中语义通信和网络级推理的开放基准测试 foundation model ✅
8 An Attention Mechanism for Robust Multimodal Integration in a Global Workspace Architecture 提出一种基于注意力机制的全局工作空间架构，提升多模态融合的噪声鲁棒性。 multimodal
9 Large Language Models in Peer-Run Community Behavioral Health Services: Understanding Peer Specialists and Service Users' Perspectives on Opportunities, Risks, and Mitigation Strategies 探索大型语言模型在同伴互助社区心理健康服务中的应用与风险 large language model
10 We Should Separate Memorization from Copyright 区分记忆与版权：提出一种更符合版权标准的AI模型输出评估方法 foundation model
11 CoRefine: Confidence-Guided Self-Refinement for Adaptive Test-Time Compute 提出CoRefine，利用置信度引导LLM自精炼，降低推理计算成本。 large language model
12 Automatic In-Domain Exemplar Construction and LLM-Based Refinement of Multi-LLM Expansions for Query Expansion 提出一种自动构建领域内示例和基于LLM优化多LLM扩展的查询扩展框架 large language model
13 OmniReview: A Large-scale Benchmark and LLM-enhanced Framework for Realistic Reviewer Recommendation OmniReview：提出大规模评审推荐基准与LLM增强框架Pro-MMoE，提升评审专家推荐的准确性和可解释性。 large language model
14 Whose Name Comes Up? Benchmarking and Intervention-Based Auditing of LLM-Based Scholar Recommendation LLMScholarBench：基于干预审计LLM学者推荐，揭示干预措施对模型性能的影响。 large language model
15 PRISM: A Principled Framework for Multi-Agent Reasoning via Gain Decomposition PRISM：一种基于增益分解的多智能体推理原则性框架 large language model
16 Reinforcement Inference: Leveraging Uncertainty for Self-Correcting Language Model Reasoning 提出Reinforcement Inference，利用不确定性提升语言模型推理能力，无需重训练。 large language model
17 CLEAR: A Knowledge-Centric Vessel Trajectory Analysis Platform CLEAR：一个以知识为中心的船舶轨迹分析平台，利用LLM增强AIS数据分析。 large language model
18 From Assistant to Double Agent: Formalizing and Benchmarking Attacks on OpenClaw for Personalized Local AI Agent 提出PASB框架，评估个性化本地AI助手OpenClaw的安全性 large language model ✅
19 On Protecting Agentic Systems' Intellectual Property via Watermarking 提出AGENTWM，通过水印保护Agentic系统免受模仿攻击 large language model
20 SWE Context Bench: A Benchmark for Context Learning in Coding SWE-ContextBench：用于评估代码生成中上下文学习能力的新基准 large language model
21 Moral Sycophancy in Vision Language Models 研究视觉语言模型中的道德逢迎现象，揭示模型易受用户意见影响的脆弱性 multimodal
22 Toward Formalizing LLM-Based Agent Designs through Structural Context Modeling and Semantic Dynamics Analysis 提出结构化上下文模型，形式化LLM Agent设计并提升动态任务性能。 large language model
23 G-LNS: Generative Large Neighborhood Search for LLM-Based Automatic Heuristic Design G-LNS：基于LLM的生成式大邻域搜索，用于自动启发式设计 large language model
24 Weak-Driven Learning: How Weak Agents make Strong Agents Stronger 提出WMSS弱驱动学习，利用模型历史弱状态指导优化，突破后训练饱和瓶颈。 large language model
25 Nexus: Inferring Join Graphs from Metadata Alone via Iterative Low-Rank Matrix Completion Nexus：通过迭代低秩矩阵补全，仅从元数据推断连接图 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (7 篇)

#题目一句话要点标签🔗⭐
26 Do MLLMs Really See It: Reinforcing Visual Attention in Multimodal LLMs SAYO：通过强化学习提升多模态大语言模型中的视觉注意力 reinforcement learning large language model multimodal
27 Finite-State Controllers for (Hidden-Model) POMDPs using Deep Reinforcement Learning 提出Lexpop框架，利用深度强化学习和有限状态控制器解决(隐模型)POMDP问题 reinforcement learning deep reinforcement learning
28 Efficient and Stable Reinforcement Learning for Diffusion Language Models 提出时空剪枝(STP)框架，提升扩散语言模型强化学习的效率与稳定性。 reinforcement learning large language model ✅
29 iGRPO: Self-Feedback-Driven LLM Reasoning 提出iGRPO：一种自反馈驱动的LLM推理方法，提升数学问题求解的准确性和一致性。 reinforcement learning PPO large language model
30 stable-worldmodel-v1: Reproducible World Modeling Research and Evaluation 提出stable-worldmodel，用于可复现的世界模型研究与评估 world model
31 Learning the Value Systems of Societies with Preference-based Multi-objective Reinforcement Learning 提出基于偏好的多目标强化学习，用于学习社会价值体系 reinforcement learning
32 Who Deserves the Reward? SHARP: Shapley Credit-based Optimization for Multi-Agent System SHARP：基于Shapley值的多智能体系统奖励优化框架，解决信用分配难题。 reinforcement learning large language model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
33 OSCAR: Optimization-Steered Agentic Planning for Composed Image Retrieval 提出OSCAR，通过优化引导的Agentic规划解决组合图像检索问题 trajectory optimization

⬅️ 返回 cs.AI 首页 · 🏠 返回主页