cs.AI（2026-09-29）

📊 共 52 篇论文 | 🔗 12 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (27 🔗6) 支柱二：RL算法与架构 (RL & Architecture) (18 🔗6) 支柱一：机器人控制 (Robot Control) (3) 支柱七：动作重定向 (Motion Retargeting) (2) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (27 篇)

#题目一句话要点标签🔗⭐
1 Pretrain Once, Route Anywhere: Towards a Foundation Model for LLM Routing 提出RouteFM以解决LLM路由效率与质量的平衡问题 large language model foundation model ✅
2 Diagnosing and Improving Probabilistic Reasoning in Large Language Models 提出决策理论框架以改善大语言模型的概率推理能力 large language model
3 OmniVCBench: Benchmarking Evidence-Grounded Multimodal Reasoning Towards AI Virtual Cells 提出OmniVCBench以解决AIVC模型解释能力不足的问题 multimodal
4 Cross-Entropy Guided Routing in Mixture-of-Experts Large Language Models 提出基于交叉熵引导的稀疏专家路由以提升语言模型性能 large language model
5 Locating Answer-Correctness Signals in Frozen Large Language Models 提出一种新方法定位冻结大语言模型中的答案正确性信号 large language model
6 VISTA: Value-Informed Event Appraisal for Multimodal Emotion Conflict 提出VISTA以解决多模态情感冲突问题 multimodal
7 DatalogBench: Evaluating Large Language Models on Text-to-Datalog Synthesis 提出DatalogBench以评估大语言模型在文本到Datalog合成中的表现 large language model
8 Probability is Not Enough: Exploring and Counting Divergent Tokens for Reasoning Uncertainty Quantification in LLMs 提出DTC框架以解决大型语言模型推理不确定性量化问题 large language model chain-of-thought ✅
9 Do LLM Agents Execute the Plans They Declare? From Planning-Mode Declaration to Pattern-Specific Execution 提出规划作为路由的方法以解决计划执行差距问题 large language model
10 Flattening the Connectome Spectrum: A Spectral Filter for FC Induces a Pretraining Target for fMRI Encoders 提出谱滤波器以优化fMRI编码器的预训练目标 foundation model
11 HARISSA: Inference-Time Self-Checks for Efficient and Safe Local Language Model Deployment 提出HARISSA以解决本地语言模型部署中的效率与安全问题 chain-of-thought
12 KV-Kaizen: Learning Context-Adaptive Cache Compression Choices 提出KV-Kaizen以解决长文本处理中的缓存压缩问题 instruction following
13 Is manual software optimization a thing of the past? 利用大型语言模型自动优化科学软件性能 large language model
14 DIET: Deletion-response Expert Trimming for Video Diffusion Transformers 提出DIET以解决视频扩散变换器专家修剪问题 large language model
15 Retrieve, Reproduce, Reveal: Dissecting Retrieval-Augmented Software Vulnerability Detection 提出统一基准以解决RAG4SVD系统可重复性与可比性问题 large language model
16 SPLASH: Switching Parallel Layouts of Attention with Seamless Handoff for LLM Serving 提出SPLASH以解决大语言模型并行注意力布局切换问题 large language model
17 Concealing LLM-Based Multi-Agent Topology via Phantom Structure Injection 提出MIRAGE框架以解决多智能体系统拓扑推断问题 large language model
18 DScale: Scaling Block-Diffusion Speculative Decoding with Adaptive Verification 提出DScale以解决大语言模型推理效率问题 large language model
19 Complexity-Aware Evaluation of LLM Comprehension 提出复杂性感知框架以评估LLM代码理解能力 large language model
20 Routing Should Pay for Itself: Sparse Supervision for Economical LLM Routing 提出SaveRouter以解决LLM路由的经济性问题 large language model ✅
21 OptiCom : A Unified Framework for State-Conditioned Composition in LLM-Driven Optimization 提出OptiCom框架以解决LLM驱动优化中的状态依赖问题 large language model
22 CRJudgeBench: Can AI Detect Plausible but Invalid Code Reviews? 提出CRJudgeBench以解决代码审查评论可信度判断问题 large language model ✅
23 SimpleEvol: An Agent-Loop Framework for LLM-Driven Automated Heuristic Design with Minimal Human Priors 提出SimpleEvol框架以优化LLM驱动的自动启发式设计 large language model ✅
24 From Learner Behavior to Reusable Skills for Effective and Efficient Learner Simulation 提出Learner2Skill以解决学习者模拟的重构成本问题 large language model
25 Breaking the Illusion of Review Reliability under Static Evaluation: SCOPE Fuzzing for LLM-based Scientific Reviewers 提出SCOPE-Fuzzer以解决静态评估下审稿可靠性问题 large language model
26 Harness Evolution as Learning: Approximation, Generalization, and Optimization Limits of Self-Improving Personal Agents 提出自我改进个人代理的演化学习方法以解决适应性问题 large language model
27 Long-Term Memory-Guided Enhancement for Target Perception in Audio-Language Models 提出长时记忆引导音频增强以改善目标感知问题 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (18 篇)

#题目一句话要点标签🔗⭐
28 Generative Interactions: Weaving Multiparty Human Motion with Bilevel Latent Dynamics 提出BRAID模型以解决多方人类运动生成问题 latent dynamics motion generation multi-person interaction
29 Train Ahead, Distill Back: Bootstrapping On-Policy Self-Distillation for Large Language Models 提出B-OPSD以提升大语言模型的自蒸馏效果 distillation privileged information large language model
30 SIPO: Unifying Reinforcement Learning with On-Policy Self-Distillation 提出SIPO以解决稀疏奖励下的自蒸馏问题 reinforcement learning distillation large language model
31 MetaCtrl: Your Large Language Models Can Reason Better and More Concisely with a Metacognitive Controller 提出MetaCtrl以动态调节推理过程，提升大语言模型的推理准确性 reinforcement learning large language model ✅
32 Beyond a single latent space: a dual-latent world model for long-horizon planning 提出双潜在世界模型以解决长时间规划问题 world model world models representation learning ✅
33 Stochastic World Models for Verifying Vision-Based Neural Feedback Systems 提出随机世界模型以验证基于视觉的神经反馈系统 world model world models
34 Spatiotemporal Hyperedges for EEG Seizure Detection and Prediction 提出HyBrain以解决EEG癫痫检测与预测问题 Mamba spatiotemporal
35 Learning to Prove, Not Just to Answer: Reinforcement Learning from Formal Verification for Natural-Language Logical Reasoning 提出Proof-R1框架以解决自然语言逻辑推理中的验证问题 reinforcement learning large language model
36 BrainNet Studio: A Unified Toolkit for Brain Network Construction, Intelligent Analysis, and Visualization 提出BrainNet Studio以解决脑网络分析的动态建模问题 predictive model spatiotemporal large language model ✅
37 AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillation 提出AdviSD以优化前沿大语言模型的建议能力 reinforcement learning distillation
38 GLaS-JEPA: Gaussian-Regularized Speech SSL without Engineered Prediction Targets 提出GLaS-JEPA框架以简化语音自监督学习 JEPA contrastive learning
39 Beyond Prompt Count: How Data Shapes Transfer in On-Policy Distillation 提出基于数据选择的高效蒸馏方法以提升教师-学生模型间的迁移能力 teacher-student distillation
40 ARGOS: Reinforcement Learning-Driven Multidimensional Elasticity for Service Orchestration in the Computing Continuum 提出ARGOS以解决计算连续体中的服务编排问题 reinforcement learning deep reinforcement learning
41 Guide, Then Let Go: Gap-Adaptive Teacher Scheduling for Sparse-Reward Agentic RL 提出Gap-Adaptive Teacher Scheduling以解决稀疏奖励强化学习中的冷启动问题 reinforcement learning teacher-student distillation ✅
42 KUPAS MASTER: Distilling the Tacit Expertise of Master Practitioners into Agent-Ready Experience Corpora 提出KUPAS MASTER以解决专业知识转化为智能体经验的问题 distillation large language model
43 Solving Without Stopping: On-Policy Distillation at Small Scale 提出小规模蒸馏训练方法以提升模型推理能力 distillation
44 PowerMarketJax: A JAX Benchmark Suite for Multi-Agent Reinforcement Learning in Power Markets 提出PowerMarketJax以解决电力市场中的多智能体强化学习问题 reinforcement learning ✅
45 CoEM: Empowering Long-Context Reasoning with Commit-on-Evidence Memory 提出Commit-on-Evidence Memory以解决长上下文推理问题 reinforcement learning large language model ✅

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
46 Direct Experience World-Model Optimization: Learning the World Beyond Action Imitation 提出直接经验世界模型优化以提升动作模仿效果 manipulation dexterous manipulation world model
47 Learning from Shared-Control Overrides: Context-Driven Acceleration Profile Prediction for Personalized Overtaking 提出基于共享控制覆盖的个性化超车加速预测方法 shared control
48 ReMem: Rethinking Perception and Memory in Long-Context Recommendation Agents 提出ReMem以解决长上下文推荐代理中的感知与记忆问题 humanoid multimodal

🔬 支柱七：动作重定向 (Motion Retargeting) (2 篇)

#题目一句话要点标签🔗⭐
49 Co-PiLOT: Constrained Physics-Informed Latent Optimization for Target-Driven Inverse Design 提出Co-PiLOT以解决物理系统逆向设计中的优化问题 latent optimization
50 Multichannel Audio Quality Assessment: Extending Pretrained Perceptual Models to Spatial Audio 提出多通道音频质量评估方法以解决空间音频评估问题 spatial relationship

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

#题目一句话要点标签🔗⭐
51 RLX: A Unified Multi-Backend Tensor Compiler and Distributed Runtime in Rust 提出RLX以统一多后端张量编译和分布式运行时 3D gaussian splatting gaussian splatting splatting

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
52 Seek Before You Move: Evidence Seeking for Progress Grounding in Vision-Language Navigation 提出SeekVLN以解决视觉语言导航中的进度短视问题 egocentric VLN

⬅️ 返回 cs.AI 首页 · 🏠 返回主页