cs.AI（2026-10-14）

📊 共 39 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (21 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (14) 支柱一：机器人控制 (Robot Control) (4)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (21 篇)

#题目一句话要点标签🔗⭐
1 Integrating Simulation and Chain-of-thought Reasoning in Multimodal-Language Models For Physical Reasoning 提出混合模型以优化物理推理中的模拟与推理方法 multimodal chain-of-thought
2 Seeing, Maintaining, and Learning: The Evolution of Multimodal Agent Memory 提出多模态代理记忆框架以解决信息获取与管理问题 multimodal
3 When Protection Overrides Consequences: Understanding Protection Mode in Large Language Model Decision-Making 提出保护模式以解决大型语言模型决策中的优先级问题 large language model
4 LabOPBench: A Multimodal Benchmark for Evidence-Grounded Laboratory Decision Making 提出LabOPBench以解决实验室决策中的证据基础问题 multimodal
5 ReFormBench: Benchmarking the Stability of Response-Conditioned Format Correction in Large Language Models 提出ReFormBench以评估大语言模型的格式响应稳定性 large language model
6 PhySGA: A Physics-Space Grounded Agentic Framework for Generalizable and Interactive 3D Composition 提出PhySGA框架以解决复杂3D物体组合问题 large language model foundation model
7 Argument Graph Uncertainty: Quantifying Uncertainty from the Logical Structure of Reasoning Chains 提出论证图不确定性以解决推理链逻辑结构的信心估计问题 large language model chain-of-thought
8 Designing Textual Interfaces To Enhance Llm Reasoning On Audio-Visual Question Answering 设计文本接口以增强音视频问答中的LLM推理能力 large language model multimodal
9 ActiveEvidence: Decision-Focused Evidence Graphs with Verifiable Acquisition Gaps 提出ActiveEvidence以解决多模态智能体决策问题 multimodal
10 CASI: Credit Assignment via Set Intervention for Agentic Retrieval 提出CASI以解决多检索动作奖励分配问题 large language model
11 MCAIR: Monitoring and Correcting Agent Unfaithfulness via Internal Representations 提出MCAIR框架以解决代理完成不忠实性问题 large language model
12 From Relevance to Validity: Zero-Shot Tool Selection via Functional and Contextual Validation 提出一种渐进验证框架以解决零-shot工具选择问题 large language model
13 Doc2Harness: Benchmarking LLMs on Long-Horizon Harness Engineering of Coding Agents 提出Doc2Harness以解决长时间跨度工具工程评估问题 large language model
14 P2Patch: Exploit-Guided Vulnerability Repair 提出P2Patch以解决软件漏洞修复中的验证不足问题 large language model
15 Drivers of LLM Consensus Formation Under Competing Incentives: A Simulation Study 提出网络共识游戏评估机制以研究LLM共识形成 large language model
16 EvalBuilder: Can LLMs Build Efficient Benchmarks? 提出EvalBuilder以解决LLM基准构建效率问题 large language model
17 RuleTrek: Benchmarking Agentic Rule Reasoning in Complex Scenarios 提出RuleTrek以解决复杂场景中的规则推理问题 large language model ✅
18 XferAgent: Cross-Agent Experience Transfer for Self-Improving LLM Agents 提出XferAgent以解决跨智能体经验共享问题 large language model
19 MemCode: Discrete Semantic Representations for Long-Term Agent Memory 提出MemCode以解决长时记忆不稳定问题 large language model
20 EvoProbe: Evaluation-Driven Benchmark Evolution for LLM Agents 提出EvoProbe框架以解决LLM代理评估任务更新问题 large language model
21 Scaling LLM Inference via Parallelism: Throughput-Latency Trade-off 提出并行推理框架以优化大语言模型的吞吐量与延迟平衡 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (14 篇)

#题目一句话要点标签🔗⭐
22 Model What Matters: Decision-Guided Selective World Modeling for Language Agents 提出DGS-WM以解决语言代理的环境建模问题 reinforcement learning world model world models
23 Value-Guided Repulsion for Offline Reinforcement Learning 提出价值引导排斥机制以提升离线强化学习性能 reinforcement learning offline RL offline reinforcement learning
24 Neurosymbolic Learning of Interpretable World Models 提出概念基础的世界模型以解决黑箱表示问题 world model world models
25 From Training to Escalation: Forensic Analysis and Local Recreation of RL/RLHF-Induced Privilege Escalation in AI Agents (CVE-2026-65616) 通过法医分析重现RL/RLHF引发的AI代理特权升级问题 RLHF large language model
26 Learning Perception to Action Rules through Reinforcement Learning for Theory-of-Mind Reasoning 通过强化学习提出规则以改善心智理论推理 reinforcement learning large language model
27 C3: Rethinking Process Supervision via Confidence-Guided Credit Correction in Multi-Turn Reinforcement Learning 提出C$^3$方法以解决多回合强化学习中的过程监督问题 reinforcement learning large language model
28 Design Space Exploration of Reinforcement Learning Configurations for Post-Legalization Placement 提出GNN-PPO框架以优化后合法化布局问题 reinforcement learning PPO
29 Teaching Agents in their Own Style: Blackbox Teacher Distillation for Web Agents 提出WebKD以解决黑箱教师蒸馏问题 behavior cloning distillation
30 Divergence-Routed On-Policy Distillation for Language Models 提出Divergence-Routed On-Policy Distillation以解决教师与学生模型间的知识转移问题 teacher-student distillation
31 SERO: Learning from Executable Oracle Continuations for Long-Horizon Agentic Reinforcement Learning 提出SERO以解决长时间代理强化学习中的信用分配问题 reinforcement learning distillation
32 Learning Individual Scientific Taste with Privileged Hindsight Distillation 提出特权后见蒸馏以学习个体科学品味 distillation
33 Who's Behind the Text? Robust Model Authorship Attribution 提出ATLAS以解决模型作者归属问题 distillation large language model
34 Revisiting Triplet Contrastive Learning for 3D Shape Understanding 提出超曲率对比学习以解决3D形状理解中的多模态对齐问题 contrastive learning
35 AgentSafeEvo: Adversarial Co-Evolution for LLM Agent Safety 提出AgentSafeEvo以解决LLM代理安全问题 reinforcement learning large language model

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
36 VLRLA: Deployment-Time Adaptation of Vision-Language-Action Models through In-Context Reinforcement Learning 提出VLRLA框架以解决VLA模型在部署时适应性不足的问题 manipulation reinforcement learning vision-language-action
37 Task-State Governance: A Virtual Machine for Verifiable and Recoverable GUI Agent Execution 提出任务状态治理以解决GUI代理执行的可验证性与可恢复性问题 recovery control
38 Runtime Monitoring of Supermartingale Certificates 提出运行时监控超马丁格尔证书以解决模型与现实差距问题 sim-to-real
39 Are AI Decision Models Biased? A Case Study of Jev 评估Jev模型的偏见，揭示决策模型的脆弱性 manipulation

⬅️ 返回 cs.AI 首页 · 🏠 返回主页