cs.AI（2026-09-01）

📊 共 45 篇论文 | 🔗 9 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (28 🔗7) 支柱二：RL算法与架构 (RL & Architecture) (13 🔗1) 支柱一：机器人控制 (Robot Control) (2) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱八：物理动画 (Physics-based Animation) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (28 篇)

#题目一句话要点标签🔗⭐
1 Embedded Conditional Independence Tests for Large Language Model Generated Text with an Application to German Parliament Speeches 提出嵌入式条件独立性检验以分析语言模型生成文本 large language model multimodal
2 MIDR: Enrichment-Augmented Indexing for Multimodal Document Retrieval 提出MIDR以解决多模态文档检索中的表示问题 multimodal
3 Differentially Private Paired Table-Image Multimodal Synthesis 提出DP-TabImage以解决多模态数据隐私合成问题 multimodal ✅
4 A Composable Evaluation System for Reproducible Omni-Modal Foundation Model Evaluation 提出可组合评估系统以解决多模态基础模型评估问题 foundation model ✅
5 H2Table: Hierarchical Hypergraph-Enhanced Large Language Models for Complex Table Reasoning 提出H2Table以解决复杂表格推理问题 large language model ✅
6 VIBE-Bench: Evaluating Personalized Large Language Models When Profiles Don't Mean Preferences 提出VIBE-Bench以解决个性化大语言模型中的偏好推理问题 large language model
7 Latent Recurrent Thoughts: Recurrent Refinement of Proposed Latents for Reasoning with Frozen LLMs 提出潜在递归思维以提升冻结LLM的推理能力 large language model chain-of-thought
8 AgentFactory: Towards Automated Agentic System Design and Optimization 提出AgentFactory以解决自动化智能系统设计与优化问题 large language model foundation model
9 Neuro-Symbolic Geometric Abstraction (NeuSOGA): From Observations to Symbolic Mathematical Representations 提出NeuSOGA以解决观察到的几何数据转化为符号数学表示的问题 foundation model
10 Measuring consistency via ensemble margin and local prediction variability: Auditing decision systems in the presence of predictive multiplicity 提出一致性测量方法以审计决策系统中的预测多样性问题 large language model
11 Can LLMs Design Video Coding Tools? A Case Study on Planar Mode 利用大型语言模型设计视频编码工具以优化编码性能 large language model
12 When Safety Routing Breaks: Understanding Alignment Fragility under Benign Fine-Tuning 提出Fisher几何解释以解决大语言模型安全对齐脆弱性问题 large language model
13 EDGE: Error Dependency Graph-Guided Multi-Error Attribution in Multi-Agent LLM Systems 提出EDGE框架以解决多代理LLM系统中的多错误归因问题 large language model
14 Automated Event Log Generation from Unstructured Text Using Finetuned LLMs 提出基于微调大语言模型的自动事件日志生成方法以解决数据结构化问题 large language model
15 Prompt-Robust Language Models: Which Training Strategies Work? 提出针对提示敏感性问题的语言模型训练策略 large language model
16 CoBRA: Learning Tool-Use Boundaries via Counterfactual Margins 提出CoBRA框架以优化工具使用决策 large language model
17 RPCBench: A Benchmark for Proactive Premise Critique in LLM-based Recommendation 提出RPCBench以解决大语言模型推荐中的前提批判问题 large language model ✅
18 In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access? 重新设计神经反馈以评估LLMs的内在表征控制能力 large language model
19 Replacing Training with Memory: Listwise Selection for Text-to-SQL 提出无需微调的列表选择器以优化文本到SQL的查询生成 large language model
20 Automated Tree Knowledge Graph Construction using Ontology Expansion and Retrieval from Vietnamese History Textbooks 提出基于本体扩展的自动化树知识图谱构建方法以解决越南语低资源问题 large language model
21 ContextPipe: Database-Inspired Context Assembly for Long-Horizon Agents 提出ContextPipe以优化长时段智能体的上下文组装问题 large language model
22 Escaping Redundant Reasoning: Structure-Aware Search for Inference-Time LLMs 提出BASIN方法以解决推理时间LLMs的冗余推理问题 large language model ✅
23 ChatDev 2.0: A No-Code Multi-Agent Platform for Developing Everything 提出ChatDev 2.0以解决多代理系统开发的复杂性问题 large language model ✅
24 Triple-Bottom-Line Sustainability of Language Models for Edge AI: A Comparison Between SLMs and Quantized LLMs 提出全面可持续性评分以比较边缘AI中的SLMs与量化LLMs large language model
25 Predicting Program Exit Code with LLMs and Programming Language Semantics 提出程序可执行性预测方法以解决LLM理解语义不足问题 large language model ✅
26 Same Request, Different Boundary: Evaluating Cybersecurity Assistance across Conversational Contexts 提出3R-Bench以评估对话上下文中的网络安全请求处理 large language model
27 Residual Sparsification via Output Importance for Compressing Mixture-of-Experts LLMs 提出PARSER以解决Mixture-of-Experts模型压缩中的准确性问题 large language model
28 WiseSpec: Requirements-Driven Agents for Code Generation 提出WiseSpec以解决代码生成中的需求质量问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (13 篇)

#题目一句话要点标签🔗⭐
29 Towards reliable multimodal disaster severity assessment through preference optimization and explainable vision-language reasoning 提出双阶段训练框架以提高灾害严重性评估的可靠性 DPO direct preference optimization multimodal
30 CARE: Contrastive Anchor-based Rubric Evolution for Large Language Model Post-Training 提出CARE以解决动态评分标准在LLM后训练中的不足问题 reinforcement learning large language model
31 ARISE-RL: Agentic Rubric-Grounded Iterative Self-Evolution with Reinforcement Learning 提出ARISE-RL以解决开放式智能体训练中的奖励不稳定问题 reinforcement learning policy learning distillation
32 When Features Become Instances: Inverted Contrastive Learning for Unsupervised Feature Selection 提出反向对比学习以解决无监督特征选择问题 representation learning contrastive learning
33 Independent Reinforcement Learning in Discounted Markov Games 提出一种新算法以解决折扣马尔可夫博弈中的独立强化学习问题 reinforcement learning
34 Text-guided flow matching enables sample-efficient crystal structure generation 提出TFMat框架以提高晶体结构生成的样本效率 flow matching
35 Reinforcement Learning Enhanced LLM Agents for Complex Vehicle Routing Problems 提出强化学习增强的LLM代理以解决复杂车辆路径问题 reinforcement learning
36 One Policy, Any Budget: Internalizing Budget-Aware Search via Reinforcement Learning 提出AnySearch框架以解决预算感知搜索问题 reinforcement learning ✅
37 SciTrue: Reliable Scientific Claim Validation with Frontier and Open Language Models at the NTCIR SciClaimEval Task 提出SciTrue以解决科学声明验证问题 distillation multimodal
38 Towards Generalizable Visually Grounded Exploration of Household Devices 提出VGEBench以解决家用设备的通用视觉基础探索问题 imitation learning affordance
39 A Closed-Loop Evaluation of Capability Loss and Recovery in Compressed Driving Policies 提出闭环评估方法以解决压缩驾驶策略的能力损失问题 PPO distillation
40 Selective Agent Guidance via Entropy: Learning Autonomous Policies from Imperfect VLM Teachers 提出SAGE框架以从不完美的VLM教师学习自主策略 reinforcement learning distillation
41 Position Matters: Feature Inversion Attacks in ViT Split Inference with Token Reduction and Shuffling 提出空间对齐重建攻击以解决ViT分割推理中的隐私问题 masked autoencoder distillation

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
42 Dual Process Motion Planning 提出双重过程运动规划以提升机器人决策效率 motion planning
43 Polished but Unresolved: Identifying Late-Stage Pressure States in Long-Horizon Tool-Use Agents 提出Probe-Sensed Pressure Relief以解决长时间工具使用代理的压力状态问题 manipulation

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
44 SOVER: Formal Certification of Optimization Reformulations via LLM-Assisted SMT Verification 提出SOVER框架以实现优化重构的形式验证 semantic mapping semantic map large language model

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

#题目一句话要点标签🔗⭐
45 Analog-DB: An Agent-First Analog Integrated Circuit Database, From Blocks to Systems 提出Analog-DB以解决模拟集成电路设计共享难题 AMP ✅

⬅️ 返回 cs.AI 首页 · 🏠 返回主页