cs.AI（2026-03-05）

📊 共 30 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (20) 支柱二：RL算法与架构 (RL & Architecture) (9) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (20 篇)

#题目一句话要点标签🔗⭐
1 K-Gen: A Multimodal Language-Conditioned Approach for Interpretable Keypoint-Guided Trajectory Generation 提出K-Gen以解决自主驾驶轨迹生成中的多模态理解问题 large language model multimodal language conditioned
2 Differentially Private Multimodal In-Context Learning 提出DP-MTV框架，实现视觉-语言模型中多模态上下文学习的差分隐私保护。 multimodal
3 Timer-S1: A Billion-Scale Time Series Foundation Model with Serial Scaling Timer-S1：通过序列化扩展实现十亿级时间序列基础模型，显著提升预测精度。 foundation model
4 Distributed Partial Information Puzzles: Examining Common Ground Construction Under Epistemic Asymmetry 提出分布式部分信息谜题（DPIP）任务，评估AI在认知不对称下的协同能力。 large language model multimodal
5 MedCoRAG: Interpretable Hepatology Diagnosis via Hybrid Evidence Retrieval and Multispecialty Consensus MedCoRAG：通过混合证据检索和多学科共识实现可解释的肝病诊断 generalist agent large language model
6 Latent-Mark: An Audio Watermark Robust to Neural Resynthesis 提出Latent-Mark，一种对神经重合成具有鲁棒性的音频水印框架。 zero-shot transfer
7 STRUCTUREDAGENT: Planning with AND/OR Trees for Long-Horizon Web Tasks STRUCTUREDAGENT：利用AND/OR树规划长程Web任务 large language model
8 X-RAY: Mapping LLM Reasoning Capability via Formalized and Calibrated Probes X-RAY：通过形式化和校准的探针映射大型语言模型的推理能力 large language model
9 GCAgent: Enhancing Group Chat Communication through Dialogue Agents System GCAgent：通过对话Agent系统增强群聊沟通 large language model
10 Escaping the Hydrolysis Trap: An Agentic Workflow for Inverse Design of Durable Photocatalytic Covalent Organic Frameworks 提出Ara智能体工作流，加速耐用光催化共价有机框架的逆向设计。 large language model
11 Training for Technology: Adoption and Productive Use of Generative AI in Legal Analysis 用户培训提升法律分析中生成式AI的采纳和生产力 large language model
12 Retrieval-Augmented Generation with Covariate Time Series 提出RAG4CTS，解决时序RAG在复杂工业场景中数据稀疏、短时和协变量耦合的难题。 foundation model
13 Beyond Linear LLM Invocation: An Efficient and Effective Semantic Filter Paradigm 提出CSV框架，通过聚类采样投票实现亚线性LLM调用，高效语义过滤。 large language model
14 DARE: Aligning LLM Agents with the R Statistical Ecosystem via Distribution-Aware Retrieval 提出DARE，通过分布感知检索对齐LLM Agent与R统计生态系统 large language model
15 Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery 利用AI辅助发现解决理论物理学中的一个开放性难题 large language model
16 The Rise of AI in Weather and Climate Information and its Impact on Global Inequality 揭示AI在气候信息中的南北差距，呼吁数据公平与知识共建 large language model foundation model
17 Reasoning Models Struggle to Control their Chains of Thought 提出CoT-Control评估套件，评估推理模型对思维链（CoT）的可控性，发现其可控性远低于输出可控性。 chain-of-thought
18 Autonomous Algorithm Discovery for Ptychography via Evolutionary LLM Reasoning Ptychi-Evolve：利用进化LLM推理实现叠层衍射成像的自主算法发现 large language model
19 SecureRAG-RTL: A Retrieval-Augmented, Multi-Agent, Zero-Shot LLM-Driven Framework for Hardware Vulnerability Detection SecureRAG-RTL：基于检索增强的多智能体零样本LLM硬件漏洞检测框架 large language model
20 EigenData: A Self-Evolving Multi-Agent Platform for Function-Calling Data Synthesis, Auditing, and Repair EigenData：一个自进化的多智能体平台，用于函数调用数据的合成、审计和修复。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (9 篇)

#题目一句话要点标签🔗⭐
21 Design Behaviour Codes (DBCs): A Taxonomy-Driven Layered Governance Benchmark for Large Language Models 提出DBC框架，通过行为约束层提升大语言模型在推理时的安全性和合规性。 RLHF DPO large language model
22 WebFactory: Automated Compression of Foundational Language Intelligence into Grounded Web Agents WebFactory：将LLM知识自动压缩为可交互Web智能体 reinforcement learning large language model foundation model
23 VISA: Value Injection via Shielded Adaptation for Personalized LLM Alignment 提出VISA框架，通过屏蔽适应的值注入实现个性化LLM对齐。 reinforcement learning RLHF large language model
24 KARL: Knowledge Agents via Reinforcement Learning 提出KARL，通过强化学习训练企业搜索Agent，在复杂搜索任务中达到SOTA性能。 reinforcement learning
25 Bidirectional Curriculum Generation: A Multi-Agent Framework for Data-Efficient Mathematical Reasoning 提出双向课程生成框架，提升大语言模型在数学推理中的数据效率。 curriculum learning large language model
26 SCoUT: Scalable Communication via Utility-Guided Temporal Grouping in Multi-Agent Reinforcement Learning SCoUT：基于效用引导时序分组的多智能体强化学习可扩展通信方法 reinforcement learning
27 Breaking Contextual Inertia: Reinforcement Learning with Single-Turn Anchors for Stable Multi-Turn Interaction 提出RLSTA方法，利用单轮锚点强化学习，解决LLM多轮交互中的上下文惯性问题。 reinforcement learning
28 Bounded State in an Infinite Horizon: Proactive Hierarchical Memory for Ad-Hoc Recall over Streaming Dialogues 提出ProStream，解决无限对话流中Ad-Hoc记忆召回的效率与准确性难题 distillation spatiotemporal
29 TimeWarp: Evaluating Web Agents by Revisiting the Past TimeWarp：通过回溯历史评估Web代理的泛化能力 behavior cloning distillation

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
30 EchoGuard: An Agentic Framework with Knowledge-Graph Memory for Detecting Manipulative Communication in Longitudinal Dialogue EchoGuard：利用知识图谱记忆检测对话中操纵性沟通的Agent框架 manipulation

⬅️ 返回 cs.AI 首页 · 🏠 返回主页