cs.AI（2026-03-12）

📊 共 29 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (22) 支柱二：RL算法与架构 (RL & Architecture) (5) 支柱一：机器人控制 (Robot Control) (1) 支柱七：动作重定向 (Motion Retargeting) (1)

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

#题目一句话要点标签🔗⭐
1 XSkill: Continual Learning from Experience and Skills in Multimodal Agents XSkill：提出一种基于经验和技能的持续学习框架，提升多模态Agent的工具使用效率。 multimodal
2 TopoBench: Benchmarking LLMs on Hard Topological Reasoning TopoBench：用于评估LLM在复杂拓扑推理能力上的基准测试 large language model chain-of-thought
3 Governing Evolving Memory in LLM Agents: Risks, Mechanisms, and the Stability and Safety Governed Memory (SSGM) Framework 提出SSGM框架，用于解决LLM Agent动态记忆中的风险、语义漂移和隐私问题 large language model multimodal
4 Explicit Logic Channel for Validation and Enhancement of MLLMs on Zero-Shot Tasks 提出显式逻辑通道以验证和增强MLLM在零样本任务中的表现 large language model multimodal
5 LLMs can construct powerful representations and streamline sample-efficient supervised learning 利用LLM构建强大表征，简化样本高效的监督学习 foundation model multimodal
6 Prototype-Based Knowledge Guidance for Fine-Grained Structured Radiology Reporting ProtoSR：利用原型知识指导细粒度结构化放射报告生成，提升图像理解能力。 multimodal
7 Human-Centred LLM Privacy Audits: Findings and Frictions 提出LMP2工具，研究LLM对个人信息的关联及用户隐私感知，揭示生成式AI评估困境。 large language model
8 Cascade: Composing Software-Hardware Attack Gadgets for Adversarial Threat Amplification in Compound AI Systems Cascade：构建软硬件攻击工具，放大复合AI系统中的对抗威胁 large language model
9 Normative Common Ground Replication (NormCoRe): Replication-by-Translation for Studying Norms in Multi-agent AI 提出NormCoRe框架，用于在多智能体AI中研究规范，通过翻译人类实验设计实现。 foundation model
10 Understanding LLM Behavior When Encountering User-Supplied Harmful Content in Harmless Tasks 揭示大语言模型在无害任务中处理用户恶意内容时的行为 large language model
11 AdaFuse: Accelerating Dynamic Adapter Inference via Token-Level Pre-Gating and Fused Kernel Optimization AdaFuse：通过Token级预选通和融合Kernel优化加速动态Adapter推理 large language model
12 You Told Me to Do It: Measuring Instructional Text-induced Private Data Leakage in LLM Agents 揭示LLM Agent中指令文本诱导的私有数据泄露风险，提出ReadSecBench基准 instruction following
13 Automating Skill Acquisition through Large-Scale Mining of Open-Source Agentic Repositories: A Framework for Multi-Agent Procedural Knowledge Extraction 提出一种框架，通过大规模挖掘开源智能体仓库自动获取技能，用于多智能体程序知识提取。 large language model
14 Gender Bias in Generative AI-assisted Recruitment Processes 揭示GenAI招聘中性别偏见：GPT-5对意大利毕业生职业建议的性别化语言模式分析 large language model
15 Scaling Laws for Educational AI Agents 提出Agent Scaling Law，通过结构化AgentProfile提升教育AI Agent能力 large language model
16 From Control to Foresight: Simulation as a New Paradigm for Human-Agent Collaboration 提出Simulation-in-the-loop，提升人机协作中Agent的决策预见性 large language model
17 Taming OpenClaw: Security Analysis and Mitigation of Autonomous LLM Agent Threats 针对OpenClaw等自主LLM Agent，提出生命周期安全框架以分析和缓解潜在威胁 large language model
18 See, Symbolize, Act: Grounding VLMs with Spatial Representations for Better Gameplay 利用空间表征增强VLMs在交互式游戏中的表现 symbolic grounding
19 Multi-Agent Collaboration for Automated Design Exploration on High Performance Computing Systems 提出MADA：基于LLM的多智能体框架，用于高性能计算系统上的自动化设计探索。 large language model
20 Speak or Stay Silent: Context-Aware Turn-Taking in Multi-Party Dialogue 提出上下文感知轮流转换方法，提升多方对话中语音助手表现 large language model
21 Deactivating Refusal Triggers: Understanding and Mitigating Overrefusal in Safety Alignment 提出基于拒绝触发词解激活的安全对齐方法，缓解大语言模型过度拒绝问题 large language model
22 XSkill: Continual Learning from Experience and Skills in Multimodal Agents XSkill：多模态Agent中基于经验和技能的持续学习框架 multimodal

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
23 Leveraging Large Language Models and Survival Analysis for Early Prediction of Chemotherapy Outcomes 利用大语言模型和生存分析早期预测化疗结果，提升患者管理。 predictive model large language model
24 On Information Self-Locking in Reinforcement Learning for Active Reasoning of LLM agents 针对LLM智能体主动推理中的信息自锁问题，提出定向评价方法。 reinforcement learning large language model
25 Agentic AI for Embodied-enhanced Beam Prediction in Low-Altitude Economy Networks 提出基于Agentic AI的混合波束预测模型，提升低空经济网络中无人机通信的波束预测精度。 Mamba large language model multimodal
26 STAIRS-Former: Spatio-Temporal Attention with Interleaved Recursive Structure Transformer for Offline Multi-task Multi-agent Reinforcement Learning 提出STAIRS-Former，利用时空注意力机制和递归Transformer解决离线多任务多智能体强化学习问题 reinforcement learning
27 Adversarial Reinforcement Learning for Detecting False Data Injection Attacks in Vehicular Routing 提出基于对抗强化学习的车辆路径虚假数据注入攻击检测方法 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
28 LABSHIELD: A Multimodal Benchmark for Safety-Critical Reasoning and Planning in Scientific Laboratories LABSHIELD：用于科学实验室中安全关键推理和规划的多模态基准测试 manipulation large language model multimodal

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
29 Stage-Adaptive Reliability Modeling for Continuous Valence-Arousal Estimation 提出SAGE框架，通过阶段自适应的可靠性建模提升连续情感识别精度。 motion estimation multimodal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页