cs.AI（2026-09-08）

📊 共 29 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) (20 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (5) 支柱三：空间感知与语义 (Perception & Semantics) (2 🔗1) 支柱一：机器人控制 (Robot Control) (1) 支柱八：物理动画 (Physics-based Animation) (1)

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

#题目一句话要点标签🔗⭐
1 Automated Design of Inventory Policy with Large Language Models: An Exploratory Study 提出基于大语言模型的自动化库存政策设计方法 large language model
2 Answer-Distribution Trajectories: A Stochastic-Dynamics View of LLM Reasoning 提出答案分布轨迹以解决LLM推理动态分析问题 chain-of-thought
3 Vision: Data-Centric Anchoring for Robust and Interpretable Agentic AI 提出数据中心锚定以解决代理AI的鲁棒性与可解释性问题 large language model
4 Procedural Graphs: Self-Evolving Execution Structures for LLM Agents 提出程序图以解决长时间规划中的行动选择问题 large language model
5 MeClear: Cooperative Game-Theoretic Attribution and Risk-Aware Memory Clearance for Long-Horizon LLM Agents 提出MeClear以解决长时间交互中的记忆管理问题 large language model
6 SQLMorph: Query Mutation and Fine-Grained Metrics for Text-to-SQL Evaluation 提出SQLMorph以解决Text-to-SQL评估中的复杂性问题 large language model
7 It's All in the Way You Say It: The Role of Information Representation in LLM-Based Glycemic-Event Prediction 基于提示的LLM方法提升糖尿病患者血糖事件预测准确性 large language model
8 Benchmark Scores Are Pipeline-Dependent: A Reliability Audit of Cybersecurity LLM Benchmarks 审计网络安全LLM基准，揭示评分依赖于评估管道 large language model
9 Graph-Based Personalized Memory for LLM Agents: Representation, Evolution, Retrieval, and Evaluation 提出图基个性化记忆以解决LLM代理的个性化问题 large language model
10 The Unreliable Progress Bar: Can LLM Agents Reliably Report Task Progress Throughout Execution? 评估大型语言模型在任务执行中报告进度的可靠性 large language model
11 AgentGrad: Intervention-guided Prompt Optimization for Multi Agent Systems 提出AgentGrad以解决多智能体系统中的提示优化问题 large language model
12 FastE: Readout-Triggered Token Compression for LLM Embedding Inference 提出FastE以解决LLM嵌入推理中的前缀冗余问题 multimodal
13 Noise Adaptive Streaming Audio-Visual Speech Token Enhancement for Robust Full-Duplex Spoken Dialogue Models 提出AV-STE以解决全双工对话系统在噪声环境下的鲁棒性问题 multimodal
14 LEBGen: An LLM-Enhanced Bayesian Network Framework for Few-Shot Travel Survey Data Generation 提出LEBGen框架以生成少量旅行调查数据 large language model
15 MemForest: Efficient Agent Memory Management via EventTree Partitioning and Progressive Merging 提出MemForest以解决代理记忆管理效率问题 multimodal ✅
16 Evidence-Aligned Entity Verification for Hallucination Detection in Retrieval-Augmented Generation 提出证据对齐实体验证方法以解决检索增强生成中的幻觉检测问题 large language model
17 ACEA: An Adversarial Co-Evolution Arena for Head-to-Head Red-Team and Blue-Team LLM Testing 提出ACEA以解决大型语言模型的红队与蓝队测试问题 large language model
18 TTGBench: Benchmarking Topological Evolution and Semantic Drift in Text-attributed Temporal Graphs 提出TTGBench以解决文本属性时间图中的语义漂移问题 large language model
19 Does Deeper Reasoning Compromise Alignment? Revealing and Mitigating of Alignment Collapse in Large Reasoning Models 提出Alignment Loss Rate以解决深度推理中的对齐崩溃问题 chain-of-thought
20 Less Is Personal: Learning Minimal Sufficient User Profiles for Personalized Language Models 提出ENOUGH方法以解决个性化语言模型的冗余用户记录问题 large language model

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

#题目一句话要点标签🔗⭐
21 WorldAgen: Unified State-Action Prediction with Test-Time World Model Training 提出WorldAgen以解决动态环境下的视觉-语言-动作适应问题 world model world models vision-language-action
22 Bridging the Semantic-Utility Gap in Multimodal RAG via Generator-in-the-Loop Alignment 提出生成器循环对齐框架以解决多模态RAG中的语义效用差距问题 DPO direct preference optimization multimodal
23 SRPO: Setwise Relative Policy Optimization for Multi-Agent LLMs 提出SRPO以解决多智能体LLMs的联合优化问题 reinforcement learning large language model
24 Do Dynamic Routers Need Memory? HeRo: History-Aware Routing for Efficient LLM Inference 提出HeRo以解决动态路由中的记忆缺失问题 linear attention large language model
25 Inference-Time Nash Alignment 提出推理时纳什对齐方法以解决偏好微调问题 RLHF DPO

🔬 支柱三：空间感知与语义 (Perception & Semantics) (2 篇)

#题目一句话要点标签🔗⭐
26 CLAMP: Constrained Decoding for Vision-Language Embodied Planning 提出CLAMP框架以解决视觉语言体规划中的约束问题 affordance multimodal
27 SchemeArena: Factorized Stress Testing of Scheming in LLM Agents 提出SCHEMEARENA以解决LLM代理中的隐秘目标问题 affordance ✅

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
28 RevalExo: A Functional Daily-Activity Benchmark for Inertial and Visual Locomotion Mode Recognition in Older Adults and Clinical Cohorts 提出RevalExo以解决老年人和临床群体的运动模式识别问题 locomotion egocentric multimodal

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

#题目一句话要点标签🔗⭐
29 AttnCompress: Dynamic Attention-Guided Trajectory Compression for Software Engineering Agents 提出AttnCompress以解决软件工程代理的动态轨迹压缩问题 ASE

⬅️ 返回 cs.AI 首页 · 🏠 返回主页