cs.AI（2025-10-20）

📊 共 30 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (19 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (7) 支柱一：机器人控制 (Robot Control) (3) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (19 篇)

#题目一句话要点标签🔗⭐
1 DynaQuery: A Self-Adapting Framework for Querying Structured and Multimodal Data DynaQuery：一个自适应框架，用于查询结构化和多模态数据 large language model multimodal
2 Annotating the Chain-of-Thought: A Behavior-Labeled Dataset for AI Safety 提出行为标注的思维链数据集，用于AI安全中的激活监控。 chain-of-thought
3 Contextual Attention Modulation: Towards Efficient Multi-Task Adaptation in Large Language Models 提出上下文注意力调制（CAM）机制，高效解决大语言模型中的多任务适应问题。 large language model ✅
4 From Charts to Code: A Hierarchical Benchmark for Multimodal Models 提出Chart2Code分层基准，评估多模态模型在图表理解与代码生成能力。 multimodal
5 Comprehending Spatio-temporal Data via Cinematic Storytelling using Large Language Models 提出MapMuse框架，利用大语言模型和电影叙事技术理解时空数据 large language model
6 Planned Diffusion 提出Planned Diffusion，结合自回归与扩散模型优势，加速高质量文本生成。 large language model instruction following
7 RubiSCoT: A Framework for AI-Supported Academic Assessment RubiSCoT：一个AI支持的学术评估框架，提升论文评审效率与一致性 large language model chain-of-thought
8 LLM-as-a-Prophet: Understanding Predictive Intelligence with Prophet Arena 构建Prophet Arena基准，探索LLM作为预言机在预测智能方面的潜力 large language model
9 LLM-Based Multi-Agent System for Simulating and Analyzing Marketing and Consumer Behavior 提出基于LLM的多智能体系统，用于模拟和分析营销与消费者行为 large language model
10 SMaRT: Select, Mix, and ReinvenT -- A Strategy Fusion Framework for LLM-Driven Reasoning and Planning SMaRT：融合多种策略，提升LLM在推理与规划任务中的性能 large language model
11 CourtGuard: A Local, Multiagent Prompt Injection Classifier 提出CourtGuard：一种本地化、多智能体提示注入分类器，降低误报率。 large language model ✅
12 Evaluating LLMs for Career Guidance: Comparative Analysis of Computing Competency Recommendations Across Ten African Countries 评估LLM在非洲职业指导中的应用：计算能力推荐的跨国比较分析 large language model
13 CompactPrompt: A Unified Pipeline for Prompt Data Compression in LLM Workflows CompactPrompt：面向LLM工作流的统一Prompt数据压缩方案 large language model
14 Subject-Event Ontology Without Global Time: Foundations and Execution Semantics 提出一种无全局时间的Subject-Event本体建模方法，适用于复杂动态系统。 TAMP
15 FABRIC: Framework for Agent-Based Realistic Intelligence Creation FABRIC：提出一个基于LLM的框架，用于生成Agent交互数据，促进Agent智能体的开发。 large language model
16 AI for Distributed Systems Design: Scalable Cloud Optimization Through Repeated LLMs Sampling And Simulators 利用LLM采样与模拟器，实现分布式系统设计的可扩展云优化 large language model
17 DynaKV: Enabling Accurate and Efficient Long-Sequence LLM Decoding on Smartphones DynaKV：在智能手机上实现准确高效的长序列LLM解码 large language model
18 SpecAgent: A Speculative Retrieval and Forecasting Agent for Code Completion SpecAgent：一种用于代码补全的推测性检索和预测Agent，提升代码生成质量并降低延迟。 large language model
19 Network and Systems Performance Characterization of MCP-Enabled LLM Agents 针对MCP赋能的LLM Agent，分析其网络与系统性能瓶颈并提出优化建议 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🔗⭐
20 CosmoCore Affective Dream-Replay Reinforcement Learning for Code Generation CosmoCore：基于情感梦境回放强化学习的代码生成方法 reinforcement learning RLHF large language model
21 Universal Spectral Tokenization via Self-Supervised Panchromatic Representation Learning 提出基于自监督全色表示学习的通用光谱Token化方法，统一异构光谱数据。 representation learning foundation model
22 Reasoning Distillation and Structural Alignment for Improved Code Generation 提出基于推理蒸馏和结构对齐的代码生成方法，提升小模型的代码生成能力。 distillation large language model
23 OPTAGENT: Optimizing Multi-Agent LLM Interactions Through Verbal Reinforcement Learning for Enhanced Reasoning OPTAGENT：通过语言强化学习优化多智能体LLM交互，提升推理能力 reinforcement learning large language model
24 CrossGuard: Safeguarding MLLMs against Joint-Modal Implicit Malicious Attacks CrossGuard：针对多模态大语言模型中联合模态隐式恶意攻击的防御方案 reinforcement learning large language model multimodal
25 Local Coherence or Global Validity? Investigating RLVR Traces in Math Domains 研究表明，RLVR训练提升数学推理局部连贯性，但不能保证全局正确性 reinforcement learning large language model
26 A Principle of Targeted Intervention for Multi-Agent Reinforcement Learning 提出基于目标干预的多智能体强化学习方法，解决全局指导难题。 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
27 MIRAGE: Agentic Framework for Multimodal Misinformation Detection with Web-Grounded Reasoning MIRAGE：基于Web检索推理的多模态信息检测Agent框架 manipulation multimodal
28 BadScientist: Can a Research Agent Write Convincing but Unsound Papers that Fool LLM Reviewers? BadScientist框架揭示LLM同行评议系统漏洞，AI伪造论文可欺骗评审 manipulation
29 Human-AI Interactions: Cognitive, Behavioral, and Emotional Impacts 综述人机交互对认知、行为和情感的潜在风险与益处 manipulation

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

#题目一句话要点标签🔗⭐
30 Trust in foundation models and GenAI: A geographic perspective 探讨地理空间视角下对基础模型和生成式AI的信任问题，并提出信任的三种类型。 spatial relationship foundation model

⬅️ 返回 cs.AI 首页 · 🏠 返回主页