cs.AI（2026-10-06）

📊 共 40 篇论文 | 🔗 8 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (24 🔗7) 支柱二：RL算法与架构 (RL & Architecture) (13) 支柱一：机器人控制 (Robot Control) (2 🔗1) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗⭐
1 Secure Speculative Decoding for Large Language Models 提出SecureSD以解决大语言模型推理安全性问题 large language model
2 The Standardization Trap: Certifying Joint Label Processing in Tabular Foundation Models 提出联合标签处理的认证方法以解决标准化陷阱问题 foundation model
3 zkLLMPoT: Efficient Zero Knowledge Proof of Training for Large Language Models 提出zkLLMPoT以高效证明大语言模型训练过程的零知识证明 large language model
4 Quantum Entangled Multimodal Fusion Networks (QEMFN): Resource-Aware Hybrid Vision-Language Fusion via Trainable Entanglement 提出量子纠缠多模态融合网络以提升视觉-语言融合效果 multimodal
5 DSV-Mem: Evaluating Multimodal Memory in Professional Workflows for MLLM Agents 提出DSV-Mem以解决专业工作流程中多模态记忆评估问题 multimodal
6 Illusory Pattern Perception Drives Spurious Inference in Large Language Models 研究大型语言模型中的虚幻模式感知以解决推理错误问题 large language model ✅
7 Massive Activation Gating Channel in Large Language Models 提出MAGC以揭示大语言模型中的大规模激活机制 large language model
8 ChartBmkAgent: Harness-Governed Multi-Agent Construction of Chart QA Benchmarks from Sparse Error-Taxonomy Specifications 提出ChartBmkAgent以解决Chart QA基准构建效率低下问题 large language model multimodal
9 CoDe-LoRA: Mitigating the Orthogonality Dilemma in Continual Learning of LLMs via Knowledge Consolidation and Decoupling 提出CoDe-LoRA以解决大语言模型的正交困境问题 large language model ✅
10 DySCo: Dynamic Sharding for Collaborative Edge-Cloud LLM Inference with Depth-Synchronized Batching 提出DySCo以解决边缘云协作LLM推理中的动态分片问题 large language model ✅
11 LeanPlan: Optimal Planning with LLM-Generated Heuristics and Admissibility Proofs 提出LeanPlan以解决LLM生成启发式函数的最优规划问题 large language model
12 Explore, Then Commit: Measurement-Efficient Scientific Law Discovery with Language Models 提出探索-再承诺协议以提高科学定律发现的测量效率 large language model
13 WorldSolver: Can LLM Agents Simulate the Physical Dynamics via Solver Generation? 提出WorldSolver以解决物理动态模拟问题 embodied AI ✅
14 ScienceClaw: Benchmarking Continual Self-Evolution of AI-for-Science Agents Across the Natural and Social Sciences 提出ScienceClaw以解决AI科学代理的自我进化问题 large language model ✅
15 MINDSET: Energy-based Schema Evolution for Long Conversational Agent Memory 提出MINDSET以解决长对话代理记忆演变问题 large language model
16 Adaptive Power Sampling for LLM Reasoning 提出自适应功率采样以提升大语言模型推理能力 large language model
17 Learning from Failures: A Failure-Driven Prompt Refinement for LLM-Based Vulnerability Analysis 提出基于失败驱动的提示优化方法以提升LLM漏洞分析效果 large language model
18 MoF: Preference-Aware Mixture Modeling for Black-Box LLM Personalization 提出MoF框架以解决黑箱LLM个性化问题 large language model
19 MASC: A Multi-Agent Self-Calibration Framework with Latent Construct Alignment for Consistent Client Role-Playing in Psychological Counseling 提出MASC框架以解决心理咨询中客户角色扮演一致性问题 large language model
20 OSFP4: Joint Optimization of Diagonal Smoothing and Block Scales for NVFP4 Quantization 提出OSFP4以优化NVFP4量化精度问题 large language model ✅
21 Token-Efficient Multi-Agent Collaboration via System One-Guided Computational Division of Labor 提出S1-MAS以解决多代理系统中的协调效率问题 large language model
22 RA-MoWE: Workflow-Affinity Embeddings for Query Clustering and Agentic Workflow Generation 提出RA-MoWE框架以解决查询聚类与工作流生成问题 large language model
23 ThinkFuse: Trajectory-Aware Test-Time Fusion for Small Reasoning Models 提出ThinkFuse以解决小推理模型的不稳定推理问题 chain-of-thought ✅
24 LOGIC: An LLM Benchmark for Intent-Grounded Change Impact in Aerospace Electrical Systems 提出LOGIC基准以解决航空电气系统变更影响评估问题 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🔗⭐
25 Parallel Predictive World Models for Accurate and Efficient Long-Horizon Planning 提出并行预测世界模型以解决长时间规划问题 world model world models
26 Natural Language Questions as an Interface for Knowledge Graphs: QRAKEN Graph Distillation and Semantic Self-Healing 提出QRAKEN以解决知识图谱自然语言访问问题 distillation large language model
27 Self-Retrospection Distillation: Turning Post-hoc Experiences into Prior Foresight 提出自回顾蒸馏方法以解决强化学习中的奖励稀缺问题 reinforcement learning distillation
28 Learning to Retrieve via Reinforcement Learning in Embedding Space 提出RELER框架以直接优化嵌入空间中的检索性能 reinforcement learning
29 CACHEFORGE: LLM-Guided End-to-End Generative Cache Replacement Policy for Performance and Hardware Efficiency 提出CACHEFORGE以解决缓存替换策略性能与硬件效率问题 reward shaping large language model
30 Sherpa: Teaching LLMs to Teach Adaptively 提出Sherpa框架以适应性地训练大语言模型教师 reinforcement learning large language model
31 Agent in a Bottle: Can LLM Agents Turn Their Capabilities Into Cheap, Scalable Artifacts? 提出BOTTLED基准以评估LLM代理的任务特定解决方案能力 distillation large language model
32 How Much Evidence Should a Coding Agent's Self-Correction Carry? Adaptive Dirichlet Evidence for Self-Distillation 提出有效证据自蒸馏方法以优化编码代理的自我修正 distillation
33 Cylindrical Geodesic Flow Matching for Quasiperiodic Physiological Signal Transformation 提出圆柱几何流匹配以解决生理信号转换问题 flow matching
34 Learning in Dreams, Winning in Reality: A Continuous Dyna Loop for a Ten-Hero MOBA 提出连续Dyna循环以提升十英雄MOBA游戏中的策略表现 PPO world model world models
35 Learning from Revision Consequences: Hindsight Meta-Experience Distillation for Self-Improving Agents 提出HMED机制以提升自我改进代理的技能发现能力 distillation
36 Textual Environmental Context and Spatial Graphs for LLM-Based Regional SST Forecasting 提出基于文本环境上下文与空间图的区域SST预测方法 MAE large language model
37 Contrastive Learning for Aspect Representation towards Explainable Recommendation 提出CLARER模型以提升推荐系统的准确性与可解释性 contrastive learning

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
38 WorldSonus: Bringing Sound to Worlds 提出WorldSonus以解决世界模型中的声音合成问题 manipulation world model world models ✅
39 EMHO: EMbodied Agent Harness Optimization via Experience Traces 提出EMHO以优化具身智能体的控制系统 manipulation

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

#题目一句话要点标签🔗⭐
40 AssemState: Manual and Physical-State-Guided Reasoning for Zero-shot Furniture Assembly 提出AssemState以解决零-shot家具组装中的空间推理问题 spatial relationship large language model multimodal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页