cs.AI（2026-08-10）

📊 共 38 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (22 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (11 🔗1) 支柱一：机器人控制 (Robot Control) (3) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱七：动作重定向 (Motion Retargeting) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (22 篇)

#题目一句话要点标签🔗⭐
1 MMArch: Benchmarking Multimodal Reasoning Grounded in Architectural Evidence 提出MMArch基准以解决多模态推理在建筑工程中的挑战 large language model multimodal ✅
2 GeoPhysAdapter: Scale-Matched Geophysical Adaptation for Cross-Domain Landslide Mapping with Vision Foundation Models 提出GeoPhysAdapter以解决跨域滑坡映射中的错误警报问题 foundation model ✅
3 SafeSceneReason: A Multimodal Reasoning Benchmark Connecting Industrial Hazards with Accident Knowledge 提出SafeSceneReason以解决工业安全推理问题 multimodal
4 MELLON - Multimodal Enhanced LLM for Online Navigation 提出MELLON以提升在线导航任务的多模态推理能力 multimodal
5 CircuitReason-1k: Benchmarking Long-Horizon Visual-to-Symbolic Reasoning inElectrical Circuits 提出CircuitReason-1k基准以评估电路的视觉到符号推理能力 large language model multimodal
6 Hallucination-Free GUI Grounding via Regression-Free Layout-Aware Matching 提出无回归布局感知匹配以解决GUI定位问题 large language model multimodal
7 CADEngBench: It Looks Like CAD, but Does It Work? Evaluating Parametric Design, Assembly Reasoning, and Physics Simulation 提出CADEngBench以评估CAD模型的工程行为 multimodal
8 MoRSE: Task-Oriented Multi-Agent System with Mixture of Role-Subtask Experts 提出MoRSE以解决多智能体系统中的任务细分与角色适应问题 large language model
9 GENCO - A Unified Neural Solver Embedded in a Development Framework for Steady-State Grid Analysis 提出GENCO以解决稳态电网分析中的多任务问题 foundation model
10 Avalon-ToM-Bench: Evaluating Fine-Grained Theory of Mind via Asymmetric Game Mechanics 提出Avalon-ToM-Bench以评估细粒度的心智理论 chain-of-thought
11 TSPORec: Token Selection via Preference Optimization for LLM-Based Sequential Recommendation 提出TSPORec以解决LLM推荐系统中的信息损失问题 large language model ✅
12 From Sweep to Seam: Interleaved Cross-Block Post-Training Quantization 提出交错跨块后训练量化方法以提升模型压缩效果 large language model
13 OpenLoopEvolve: A Verifiable Self-Evolution Framework for Loop Policies in Long-Horizon Complex Tasks 提出OpenLoopEvolve框架以解决长时间复杂任务中的自我演化问题 large language model
14 P$^{3}$: Joint Program-and-Proof Planning for Verified Code Generation 提出P$^{3}$以解决验证代码生成中的效率与有效性问题 large language model
15 SafeQL: Search-based Refinement for Safe and Efficient LLM-based Text-to-SQL 提出SafeQL以解决LLM文本到SQL生成中的安全性与效率问题 large language model
16 Governing the KV Cache: Preventing Timing Side-Channel Leakage in Multi-Tenant LLM Inference 提出KVGov以解决多租户LLM推理中的时序侧信道泄露问题 large language model
17 Agentic Router: An Execution-Grounded Continual Learning Approach With Memory 提出执行基础的持续学习方法以提升CLI操作的成功率 large language model
18 From Relevance to Execution Utility: Reward-Aware Dynamic Execution Gating for Skill-Based LLM Agents 提出奖励感知动态执行门以优化技能基础LLM代理的执行效率 large language model
19 CIDER: A Dataset of Contextual Disclosure Boundaries for Privacy Preference Alignment 提出CIDER数据集以解决隐私偏好对齐问题 large language model
20 ChronoState: Hidden Elapsed-Time Conditioning for Temporal-State Action Selection in Frozen-Backbone Language Models 提出ChronoState以解决语言模型中的时间状态选择问题 TAMP
21 RAVEN-Eval: Rubric-Guided Automatic Evaluation for AI Video Generation Models Based on LMM Preference Judgement 提出RAVEN-Eval以解决AI视频生成模型评估难题 instruction following
22 Who Bridges Safety? Identifying and Targeting Cross-Lingual Shared Safety Pathways 提出跨层安全路径识别方法以解决多语言安全差距问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (11 篇)

#题目一句话要点标签🔗⭐
23 Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models 提出模型发现代理以解决数据效率低下的问题 world model world models large language model
24 Bidirectional Context Self-Distillation for Reinforcement Learning of Skill-Based LLM Agents 提出双向上下文自蒸馏框架以提升技能型LLM代理的强化学习效果 reinforcement learning distillation large language model
25 Different Feedback, Different Updates: Selective Self-Learning from User Interactions for Large Language Models 提出SLIFT框架以优化大语言模型的用户反馈学习 distillation large language model
26 verdi: retrieval is not transfer for continual world model optimization 提出VERDI框架以解决世界模型优化中的知识转移问题 world model world models
27 Listen, See and Track: Spatio-Temporal Audio-Visual Sound Event Reasoning for Omni-Modal Language Models 提出ST-Omni-R1以解决动态声音源理解问题 reinforcement learning curriculum learning motion representation
28 Stealing Reasoning Traces from Proprietary LLM APIs 提出一种方法以从专有LLM API中提取推理痕迹 distillation large language model chain-of-thought
29 Mismatch Matters: On-Policy Distillation Beyond Token Agreement 提出TIDE以解决长文本生成中的教师-学生不匹配问题 teacher-student distillation reward shaping ✅
30 Coupled Graph--Policy Distillation for Personalized Medication Safety in Older Adults with Multimorbidity 提出ATLAS框架以提升老年多病患者的用药安全性 distillation large language model
31 RISE-RL: Rubric-Informed Selective Exploration for Open-Ended Reinforcement Learning 提出RISE-RL以解决开放式强化学习中的多维标准对齐问题 reinforcement learning large language model
32 Motif 3: Technical Report 提出Motif 3以提升语言模型的推理与长文本理解能力 reinforcement learning distillation instruction following
33 Privileged Likelihood Is Not Automatically Value: Three Checks for Token Credit in On-Policy Self-Distillation 提出三项检查以验证自蒸馏中的令牌信用问题 distillation

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
34 Control-Oriented Scenario Tree Construction through Reinforcement Learning 提出基于强化学习的场景树构建方法以优化控制性能 MPC model predictive control reinforcement learning
35 MADBench: A Benchmark for Modality-Aware Audio Deepfake Detection 提出MADBench以解决音频深伪检测中的模态区分问题 manipulation large language model multimodal
36 MusicLayout: Explicit Structural Planning for Controllable Text-to-Music Generation 提出MusicLayout以解决文本到音乐生成中的结构控制问题 manipulation

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

#题目一句话要点标签🔗⭐
37 GLocFM: A Geometry-Aware Foundation Model for 3D Indoor Wireless Localization 提出GLocFM以解决3D室内无线定位中的几何信息利用不足问题 NeRF foundation model

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

#题目一句话要点标签🔗⭐
38 Emotion2Skill: Model-Internal Emotion Signals for Adaptive Skill Selection and Evolution 提出Emotion2Skill框架以解决LLM代理技能选择问题 motion representation ✅

⬅️ 返回 cs.AI 首页 · 🏠 返回主页