cs.AI（2026-06-22）

📊 共 29 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (20 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (5) 支柱四：生成式动作 (Generative Motion) (1 🔗1) 支柱一：机器人控制 (Robot Control) (1) 支柱八：物理动画 (Physics-based Animation) (1) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (20 篇)

#题目一句话要点标签🔗⭐
1 Kamera: Unified Position-Invariant Multimodal KV Cache for Training-Free Reuse 提出Kamera以解决多模态缓存重编码问题 multimodal
2 TailorMind: Towards Preference-Aligned Multimodal Content Generation 提出TailorMind以解决个性化多模态内容生成问题 multimodal ✅
3 POTracker: Optimizing Large Language Models for Standard-Compliant Power Outage Report Generation 提出POTracker以解决电力中断报告生成的标准化问题 large language model
4 Abstract representational geometry supports inference in large language models 提出抽象表征几何以支持大型语言模型推理 large language model
5 TTFT-Aware Graph Chain-of-Thought:Distance-Indexed Neural A* for Low-Hallucination Multi-Hop Medical Reasoning 提出TTFT感知图链式思维以解决医疗推理中的幻觉问题 chain-of-thought
6 When Preferences Fail to Become Incentives: A Utility-Behavior Gap in Large Language Models 探讨偏好与激励之间的差距以提升大语言模型行为理解 large language model
7 Graph-Enhanced Large Language Models for Spatial Search 提出图增强大语言模型以解决空间推理问题 large language model
8 Measuring Behavior Portability in Large Language Models 提出行为可移植性测量框架以解决LLM决策环境问题 large language model
9 The Origins of Stochasticity: Comprehensive Investigations on Uncertainty Quantification for Large Language Models 提出细粒度不确定性分类以解决大语言模型的预测可信度问题 large language model
10 ENVS: Environment-Native Verified Search for Long-Horizon GUI Agents 提出ENVS以解决长时间序列GUI任务的有效搜索问题 multimodal
11 Causal Discovery in the Era of Agents 提出代理辅助的因果发现方法以解决数据与假设的混淆问题 large language model
12 LLM-Aided A* Search in Non-Geometric Network Graphs 提出LLM辅助的A*搜索算法以解决非几何网络图中的最短路径问题 large language model
13 AI Exposure Scores: what they measure, what they miss, and what comes next 提出AI曝光评分以解决政策与研究之间的协调问题 large language model
14 The Topology of Ill-Posed Questions: Persistent Homology for Detection and Steering in LLMs 提出持久同调方法以解决大型语言模型中的不适定问题 large language model
15 Detecting Malicious Agent Skills in the Wild using Attention 提出Locate-and-Judge以检测恶意代理技能 instruction following
16 GIF: Locally Sound Geometric Information Flow Control for LLMs 提出GIF框架以解决大语言模型的信息流控制问题 large language model
17 The Model as One Rater Among Several: Measuring Political Positions in Data-Sparse Regions with a Language-Model Panel 提出语言模型面板以解决数据稀缺地区的政治立场测量问题 large language model
18 A Stackelberg Framework for Resource-Aware LLM Agents: Learning, Repair, and Conditional Guarantees 提出基于Stackelberg框架的资源感知LLM代理以优化资源分配 large language model
19 From Fragments to Paths: Task-Level Context Recovery for Large Industrial Codebases 提出DeepDiscovery以解决大型工业代码库的任务级上下文恢复问题 large language model
20 Agent-as-a-Router: Agentic Model Routing for Coding Tasks 提出Agent-as-a-Router框架以优化多模型任务路由 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
21 VeriEvol: Scaling Multimodal Mathematical Reasoning via Verifiable Evol-Instruct 提出VeriEvol以解决可验证的多模态数学推理问题 reinforcement learning multimodal
22 SQLConductor: Search-to-Policy Learning for Step-wise Text-to-SQL Orchestration 提出SQLConductor以解决复杂数据库环境下的Text-to-SQL问题 reinforcement learning policy learning
23 Finding the Evidence: Discovering Decision-Supporting Tokens for On-Policy Reasoning Distillation 提出DEAR方法以解决决策支持信号转移不足问题 teacher-student distillation
24 A Formula-Driven Survey and Research Agenda for On-Policy Distillation 提出基于公式的反馈更新框架以优化在政策蒸馏中的效果 teacher-student distillation
25 SPIRAL: Learning to Search and Aggregate 提出SPIRAL框架以提升语言模型推理能力 reinforcement learning chain-of-thought

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
26 Text Dictates, Music Decorates: Energy-based Attention for Editable Dance Motion Generation 提出STREAM以解决舞蹈动作生成中的语义控制问题 motion generation ✅

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
27 Teaching LLMs String Matching, Backtracking, and Error Recovery to Deduce Bases and Truth Tables for the Combinatorially Exploding Bit Manipulation Puzzles 提出一种新方法解决大规模位操作谜题中的逻辑推理问题 manipulation large language model

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

#题目一句话要点标签🔗⭐
28 Evolutionary Optimization Reveals Structural Constraints on Reservoir Architecture for Spatiotemporal Chaos 通过进化优化揭示水库架构的结构约束以应对时空混沌 spatiotemporal

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

#题目一句话要点标签🔗⭐
29 Towards a Bathroom-Centered Human-Building Digital Twin Framework for Indoor Safety Analysis 提出以浴室为中心的人-建筑数字双胞胎框架以提升室内安全性 human motion

⬅️ 返回 cs.AI 首页 · 🏠 返回主页