cs.AI（2026-02-13）

📊 共 27 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) (16 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗2) 支柱一：机器人控制 (Robot Control) (3) 支柱六：视频提取与匹配 (Video Extraction) (1) 支柱四：生成式动作 (Generative Motion) (1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (16 篇)

#题目一句话要点标签🔗⭐
1 How Multimodal Large Language Models Support Access to Visual Information: A Diary Study With Blind and Low Vision People 研究多模态大语言模型如何辅助视障人士获取视觉信息，揭示其在实际应用中的挑战与机遇。 large language model multimodal
2 BrowseComp-$V^3$: A Visual, Vertical, and Verifiable Benchmark for Multimodal Browsing Agents 提出BrowseComp-$V^3$多模态浏览Agent基准，解决现有基准在复杂性、可访问性和评估粒度上的局限性。 large language model multimodal
3 TriGen: NPU Architecture for End-to-End Acceleration of Large Language Models based on SW-HW Co-Design TriGen：基于软硬件协同设计的端到端大语言模型加速NPU架构 large language model
4 Assessing Spear-Phishing Website Generation in Large Language Model Coding Agents 评估大型语言模型编码智能体生成鱼叉式网络钓鱼网站的能力 large language model
5 RQ-GMM: Residual Quantized Gaussian Mixture Model for Multimodal Semantic Discretization in CTR Prediction 提出RQ-GMM，用于CTR预测中多模态语义离散化，提升点击率。 multimodal
6 Artic: AI-oriented Real-time Communication for MLLM Video Assistant Artic：面向MLLM视频助手的AI实时通信框架，提升准确率并降低延迟 large language model multimodal ✅
7 Protect$^*$: Steerable Retrosynthesis through Neuro-Symbolic State Encoding Protect$^*$: 提出神经符号框架，通过可控的逆合成分析指导LLM生成化学反应路径。 large language model
8 AI Agents for Inventory Control: Human-LLM-OR Complementarity 提出人-LLM-OR协同的库存控制AI Agent，提升复杂场景下的决策性能 large language model
9 Arming Data Agents with Tribal Knowledge Tk-Boost：利用部落知识增强NL2SQL数据代理，提升查询准确性 large language model
10 Asynchronous Verified Semantic Caching for Tiered LLM Architectures Krites：异步验证语义缓存，提升分层LLM架构静态缓存覆盖率 large language model
11 Buy versus Build an LLM: A Decision Framework for Governments 提出决策框架以帮助政府选择LLM的购买或构建策略 large language model
12 G2CP: A Graph-Grounded Communication Protocol for Verifiable and Efficient Multi-Agent Reasoning 提出G2CP图谱通信协议，解决多智能体系统中的语义漂移和幻觉问题 large language model
13 Knowledge-Based Design Requirements for Generative Social Robots in Higher Education 针对高等教育中生成式社交机器人，提出基于知识的设计需求框架 large language model
14 Think Fast and Slow: Step-Level Cognitive Depth Adaptation for LLM Agents CogRouter：为LLM Agent设计认知深度自适应框架，提升效率与性能。 large language model
15 TensorCommitments: A Lightweight Verifiable Inference for Language Models TensorCommitments：一种轻量级的语言模型可验证推理方案 large language model
16 GeoAgent: Learning to Geolocate Everywhere with Reinforced Geographic Characteristics GeoAgent：通过强化地理特征学习在任意地点进行地理定位 chain-of-thought

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

#题目一句话要点标签🔗⭐
17 To Mix or To Merge: Toward Multi-Domain Reinforcement Learning for Large Language Models M2RL：探索混合训练与模型合并在多领域大语言模型强化学习中的优劣 reinforcement learning large language model instruction following ✅
18 In-Context Autonomous Network Incident Response: An End-to-End Large Language Model Agent Approach 提出基于LLM的端到端Agent，用于自主网络事件响应，无需手工建模。 reinforcement learning large language model chain-of-thought
19 UBio-MolFM: A Universal Molecular Foundation Model for Bio-Systems UBio-MolFM：用于生物系统的通用分子基础模型，实现量子精度与生物尺度的统一。 curriculum learning foundation model
20 On-Policy Supervised Fine-Tuning for Efficient Reasoning 提出On-Policy SFT，通过监督微调提升大推理模型的效率与准确率。 reinforcement learning chain-of-thought ✅
21 Information-theoretic analysis of world models in optimal reward maximizers 量化最优策略所需的世界模型信息量下界，揭示智能行为的内在表征需求 world model

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
22 Backdooring Bias in Large Language Models 研究表明，白盒攻击下，语义触发后门更易诱导大语言模型的负面偏见。 manipulation large language model
23 Never say never: Exploring the effects of available knowledge on agent persuasiveness in controlled physiotherapy motivation dialogues 通过知识配置提升生成式社交Agent在理疗激励对话中的说服力 manipulation
24 WebClipper: Efficient Evolution of Web Agents with Graph-based Trajectory Pruning WebClipper：基于图剪枝的高效Web Agent进化框架 trajectory optimization

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
25 "Not Human, Funnier": How Machine Identity Shapes Humor Perception in Online AI Stand-up Comedy 利用机器身份进行AI脱口秀：提升在线AI喜剧的幽默感知 HuMoR

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
26 Can I Have Your Order? Monte-Carlo Tree Search for Slot Filling Ordering in Diffusion Language Models 提出McDiffuSE，利用蒙特卡洛树搜索优化扩散语言模型中的槽填充顺序，提升生成质量。 MDM

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
27 REMem: Reasoning with Episodic Memory in Language Agent REMem：提出一种基于情景记忆的语言代理推理框架，提升复杂推理能力。 spatiotemporal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页