cs.AI（2026-02-03）

📊 共 35 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) (23 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗1) 支柱一：机器人控制 (Robot Control) (3) 支柱八：物理动画 (Physics-based Animation) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (23 篇)

#题目一句话要点标签🔗⭐
1 Towards Considerate Embodied AI: Co-Designing Situated Multi-Site Healthcare Robots from Abstract Concepts to High-Fidelity Prototypes 通过共创设计，为多场景医疗机器人打造更周到的具身AI embodied AI ✅
2 Agentic Proposing: Enhancing Large Language Model Reasoning via Compositional Skill Synthesis 提出Agentic Proposing框架，通过组合技能合成高质量数据，提升大语言模型推理能力。 large language model
3 An Empirical Study of Collective Behaviors and Social Dynamics in Large Language Model Agents 研究大型语言模型Agent的集体行为与社会动态，提出CoST方法抑制有害信息发布 large language model
4 VALUEFLOW: Toward Pluralistic and Steerable Value-based Alignment in Large Language Models 提出VALUEFLOW以解决大语言模型价值对齐问题 large language model
5 De-conflating Preference and Qualification: Constrained Dual-Perspective Reasoning for Job Recommendation with Large Language Models JobRec：通过约束双视角推理，解耦偏好与资格，用于LLM驱动的职位推荐 large language model
6 Large Language Models Can Take False First Steps at Inference-time Planning 大型语言模型在推理时规划中存在虚假先验步骤问题 large language model
7 CSR-Bench: A Benchmark for Evaluating the Cross-modal Safety and Reliability of MLLMs 提出CSR-Bench，用于评估多模态大语言模型跨模态安全性和可靠性 large language model multimodal
8 Are LLMs Biased Like Humans? Causal Reasoning as a Function of Prior Knowledge, Irrelevant Information, and Reasoning Budget 评估大型语言模型的因果推理与人类偏见的关系 large language model chain-of-thought
9 Conformal Thinking: Risk Control for Reasoning on a Compute Budget 提出基于风险控制的自适应推理框架，优化大语言模型在计算预算下的推理。 large language model
10 DiscoverLLM: From Executing Intents to Discovering Them 提出DiscoverLLM框架，通过意图发现提升LLM在开放式任务中的交互性能。 large language model
11 Methods and Open Problems in Differentiable Social Choice: Learning Mechanisms, Decisions, and Alignment 综述可微社会选择方法：学习机制、决策与对齐，并提出开放性问题。 large language model
12 Persona Generators: Generating Diverse Synthetic Personas at Scale 提出Persona Generators，利用进化算法生成多样化合成角色，用于AI系统评估。 large language model
13 When Routing Collapses: On the Degenerate Convergence of LLM Routers 提出EquiRouter以解决LLM路由中的退化收敛问题，提升成本效益。 multimodal ✅
14 Ontology-to-tools compilation for executable semantic constraint enforcement in LLM agents 提出本体到工具的编译方法，用于LLM Agent中可执行的语义约束强化 large language model
15 Precision in Practice: Knowledge Guided Code Summarizing Grounded in Industrial Expectations ExpSum：结合工业期望的知识引导代码摘要生成方法 large language model
16 The Necessity of a Unified Framework for LLM-Based Agent Evaluation 提出LLM Agent统一评估框架，解决评估标准不一致问题 large language model
17 Beyond Quantity: Trajectory Diversity Scaling for Code Agents TDScaling：通过轨迹多样性提升代码智能体性能，突破数量 scaling 瓶颈 large language model
18 Internet of Agentic AI: Incentive-Compatible Distributed Teaming and Workflow 提出Internet of Agentic AI框架，实现可扩展的Agentic AI分布式协作与工作流。 large language model
19 Understanding Multi-Agent LLM Frameworks: A Unified Benchmark and Experimental Analysis 提出MAFBench，用于系统评估多智能体LLM框架架构对性能的影响 large language model
20 Digital Lifelong Learning in the Age of AI: Trends and Insights 分析AI时代终身数字学习趋势，揭示学习动机与平台优化策略 large language model
21 Risky-Bench: Probing Agentic Safety Risks under Real-World Deployment 提出Risky-Bench以解决现实环境中代理安全风险评估问题 large language model
22 MAS-ProVe: Understanding the Process Verification of Multi-Agent Systems MAS-ProVe：系统性研究多智能体系统过程验证的有效性与挑战 large language model ✅
23 RC-GRPO: Reward-Conditioned Group Relative Policy Optimization for Multi-Turn Tool Calling Agents 提出RC-GRPO，通过奖励调节提升多轮工具调用Agent的性能 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🔗⭐
24 EHRWorld: A Patient-Centric Medical World Model for Long-Horizon Clinical Trajectories EHRWorld：面向长期临床轨迹的以患者为中心的医疗世界模型 world model large language model
25 IntentRL: Training Proactive User-intent Agents for Open-ended Deep Research via Reinforcement Learning 提出IntentRL以解决深度研究代理的用户意图澄清问题 reinforcement learning large language model
26 Mitigating Conversational Inertia in Multi-Turn Agents 提出上下文偏好学习，缓解多轮Agent对话中的惯性问题 preference learning large language model
27 General Agents Contain World Models, even under Partial Observability and Stochasticity 证明通用智能体即使在部分可观测和随机环境下也包含世界模型 world model
28 STAR: Similarity-guided Teacher-Assisted Refinement for Super-Tiny Function Calling Models STAR：面向超小型函数调用模型的相似性引导的教师辅助精炼 distillation large language model
29 Distilling LLM Reasoning into Graph of Concept Predictors 提出概念预测图（GCP）框架，用于将LLM推理能力蒸馏到小型判别模型中，提升效率和可解释性。 distillation large language model ✅
30 Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration 提出Search-R2以解决搜索集成推理中的多尺度信用分配问题 reinforcement learning reward design

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
31 CRL-VLA: Continual Vision-Language-Action Learning 提出CRL-VLA框架，解决具身智能体持续视觉-语言-动作学习中的稳定性-可塑性难题 manipulation dexterous manipulation reinforcement learning
32 Risk Awareness Injection: Calibrating Vision-Language Models for Safety without Compromising Utility 提出风险感知注入（RAI），提升视觉语言模型安全性且不损失性能 manipulation large language model multimodal
33 Can LLMs Do Rocket Science? Exploring the Limits of Complex Reasoning with GTOC 12 评估LLM在复杂航天任务中的能力：GTOC 12挑战 trajectory optimization large language model

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

#题目一句话要点标签🔗⭐
34 Morphe: High-Fidelity Generative Video Streaming with Vision Foundation Model 提出Morphe以解决视频流传输中的质量与延迟问题 spatiotemporal foundation model
35 Adaptive Evidence Weighting for Audio-Spatiotemporal Fusion 提出FINCH框架以解决生物声学分类中的证据融合问题 spatiotemporal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页