cs.AI（2026-06-23）

📊 共 31 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (14 🔗1) 支柱九：具身大模型 (Embodied Foundation Models) (14 🔗2) 支柱一：机器人控制 (Robot Control) (3 🔗1)

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

#题目一句话要点标签🔗⭐
1 Scaling Laws for Task-Specific LLM Distillation 提出领域特定LLM蒸馏的缩放规律以解决压缩挑战 distillation large language model chain-of-thought
2 FlowR2A: Learning Reward-to-Action Distribution for Multimodal Driving Planning 提出FlowR2A以解决多模态驾驶规划中的奖励与动作分布问题 flow matching multimodal
3 World Models in Pieces: Structural Certification for General Agents 提出结构认证以解决通用智能体的局限性问题 world model world models
4 LaGO: Latent Action Guidance for Online Reinforcement Learning 提出LaGO框架以改善在线强化学习中的决策效率 reinforcement learning PPO large language model
5 CineCap: Structured Reasoning with Spatio-Temporal Anchors for Cinematographic Video Captioning 提出CineCap以解决电影视频字幕生成中的专业概念理解问题 reinforcement learning large language model multimodal ✅
6 An Introduction to Causal Reinforcement Learning 提出因果强化学习以解决因果推断与强化学习的结合问题 reinforcement learning policy learning imitation learning
7 ASALT: Adaptive State Alignment for Lateral Transfer in Multi-agent Reinforcement Learning 提出ASALT以解决多智能体强化学习中的状态空间不匹配问题 reinforcement learning
8 Solving Inverse Problems of Chaotic Systems with Bidirectional Conditional Flow Matching 提出双向条件流匹配以解决混沌系统逆问题 flow matching
9 Beyond U-Net: A Latent-Representation-Aligned Skip-Free Backbone for Flow-Matching Speech Enhancement 提出跳跃连接自由的编码解码骨干网络以解决语音增强问题 flow matching
10 Themis: An explainable AI-enabled framework for Reinforcement Learning with Human Feedback 提出Themis框架以解决强化学习中的人类反馈与透明性问题 reinforcement learning
11 Reinforcement Learning for Computer-Use Agents with Autonomous Evaluation 提出基于自主评估的强化学习框架以优化计算机使用代理 reinforcement learning
12 Accelerating Disaggregated RL for Visual Generative LLMs with Diffusion-Based Parallelism and Trainer-Assisted Generation 提出DigenRL框架以提升扩散生成LLMs的强化学习效率 reinforcement learning large language model
13 Real-Time Interactive Music Generation via Data-Free Streaming Consistency Distillation 提出实时互动音乐生成框架以解决生成延迟问题 distillation
14 The Geometry Behind Diffusion and Flow Matching: Gradient Flows and Geodesics in Wasserstein Space 提出基于Wasserstein空间的几何方法以统一扩散与流匹配模型 flow matching

🔬 支柱九：具身大模型 (Embodied Foundation Models) (14 篇)

#题目一句话要点标签🔗⭐
15 PHANTOM: A Large-Scale Dataset of Multimodal Adversarial Attacks for Vision-Language Models 提出PHANTOM数据集以支持视觉语言模型的对抗攻击研究 multimodal
16 Large-Language-Model Discovery of Quantum LDPC Codes through Structured Concept Evolution 提出结构化概念进化方法以发现量子LDPC码 large language model
17 Can Scale Save Us From Plasticity Loss in Large Language Models? 研究表明规模无法完全解决大语言模型的可塑性损失问题 large language model
18 A specialized reasoning large language model for accelerating rare disease diagnosis: a randomized AI physician assistance trial 提出RaDaR以加速罕见疾病诊断 large language model
19 On the Smallness of the Large Language Models Scaling Exponents 探讨大语言模型缩放指数的可持续性问题 large language model
20 Age of LLM: A Strategic 1v1 Benchmark for Reasoning, Diplomacy and Reliability of Large Language Models under Fog of War 提出Age of LLM基准以评估大语言模型在战争迷雾下的推理与外交能力 large language model
21 Navigating User Behavior toward Personalized Multimodal Generation 提出NaviGen以解决个性化多模态生成中的用户行为理解问题 multimodal ✅
22 AdversaBench: Automated LLM Red-Teaming with Multi-Judge Confirmation and Cross-Model Transferability 提出AdversaBench以解决大语言模型的对抗性评估问题 large language model instruction following ✅
23 When Helpfulness Overrides Causal Caution: Context-Dependent Suppression and Recovery in LLMs 研究LLMs在决策支持中的因果谨慎抑制与恢复机制 large language model
24 LemonHarness Technical Report 提出LemonHarness以解决长任务中状态追踪困难的问题 large language model
25 Probing the Misaligned Thinking Process of Language Models 提出细粒度认知过程监测方法以检测语言模型的行为失调 large language model
26 AutoSpec: Safety Rule Evolution for LLM Agents via Inductive Logic Programming 提出AutoSpec框架以解决LLM代理的安全规则演化问题 large language model
27 Agon: An Autonomous Large-Scale Omnidisciplinary Research System Built on Prompt Economy 提出Agon以实现大规模跨学科研究的自动化 large language model
28 VeryTrace: Verifying Reasoning Traces through Compilable Formalism and Structured Verification 提出VeryTrace以解决多步推理验证问题 chain-of-thought

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
29 ReMMD: Realistic Multilingual Multi-Image Agentic Verification for Multimodal Misinformation Detection 提出ReMMD框架以解决多模态虚假信息检测问题 manipulation multimodal ✅
30 Adaptive Machine Learning Framework for UAV Trajectory Optimization in O-RAN 提出自适应机器学习框架以优化无人机轨迹 trajectory optimization
31 Red-Teaming the Agentic Red-Team 提出针对代理系统的安全分析与架构设计以解决安全漏洞问题 manipulation

⬅️ 返回 cs.AI 首页 · 🏠 返回主页