cs.LG（2026-08-27）

📊 共 27 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (14) 支柱九：具身大模型 (Embodied Foundation Models) (8) 支柱一：机器人控制 (Robot Control) (3 🔗1) 支柱五：交互与反应 (Interaction & Reaction) (1) 支柱八：物理动画 (Physics-based Animation) (1)

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

#题目一句话要点标签🔗⭐
1 Graph-Based Pseudo-multimodal Contrastive Learning for 12-Lead ECG Representations 提出图基伪多模态对比学习以解决12导联ECG分析问题 contrastive learning multimodal
2 Simple Actors and Deep Critics for Scalable Reinforcement Learning 提出LAC以解决离线强化学习中的效率问题 reinforcement learning offline RL offline reinforcement learning
3 Mitigating Strong-Modality Collapse in Multimodal Learning via Inverted Asymmetric Fusion 提出反向非对称融合以解决多模态学习中的强模态崩溃问题 distillation mutual attention multimodal
4 HALO: A Heterogeneity-Aware Language-Aligned IMU Foundation Model for Open-Set Human Activity Recognition 提出HALO模型以解决开放集人类活动识别中的异构性问题 contrastive learning foundation model
5 Diffusion Policies for Short-Horizon Planning in Robot Crowd Navigation 提出规划扩散策略优化以解决机器人群体导航问题 reinforcement learning PPO diffusion policy
6 Making Latent Evolution Explicit: Operator-Structured Transitions for World Action Models 提出LEON以解决潜在演化显式化问题 world action model world action models
7 Performance Foundations of Parallel & Distributed Reasoning Language Models 提出高效的并行与分布式推理语言模型训练方法 reinforcement learning PPO large language model
8 Reinforcement Learning-Based Control of CAV Platoon Joining Maneuvers in Mixed Traffic 提出基于强化学习的控制方法以解决混合交通中的CAV编队加入问题 reinforcement learning deep reinforcement learning DRL
9 Disentangling Optimization Scale from Preference Scale in DPO 提出中心软正则化以解决DPO中的优化规模与偏好规模纠缠问题 DPO direct preference optimization
10 A Finite Sample Analysis for Quantile Temporal Difference Learning in Distributional Reinforcement Learning 提出全局有限样本保证以优化分布式强化学习中的量化时间差学习 reinforcement learning
11 Scaling Model-Generated Distillation Data Can Make Latent Teacher Traits More Recoverable 提出通过扩展模型生成的蒸馏数据以增强教师特征的可恢复性 distillation
12 Technical Comparative Benchmarking Study: Advanced AI Hybrid Methods for Renewable Energy Farm Optimization and Forecasting 提出先进AI混合方法以优化和预测可再生能源农场 representation learning MAE
13 Diffusion Distillation for Efficient Weather Ensembles 提出扩散蒸馏方法以提高天气预报效率 distillation
14 SegBench-GC: Testing Segmentation Invariance in Multi-Step Offline Goal-Conditioned Reinforcement Learning 提出SegBench-GC以解决离线目标条件强化学习中的分段不变性问题 reinforcement learning

🔬 支柱九：具身大模型 (Embodied Foundation Models) (8 篇)

#题目一句话要点标签🔗⭐
15 Activation Outliers Matter: Robust Recovery for Quantized Multimodal LLMs 提出残差回退量化以解决多模态大语言模型的激活量化问题 large language model multimodal
16 MM-Spectrum: Multimodal Multi-spectral Molecular Structural Elucidation with a Stable MoE Framework 提出MM-Spectrum以解决多模态光谱分子结构推断问题 multimodal
17 SAGE: Variate-Wise Semantic Augmentation for Vision-Language Time Series Forecasting 提出SAGE框架以解决时间序列预测中的语义缺失问题 large language model multimodal
18 Unifying Detection and Adaptation in Task-Free Continual Learning 提出FiUni框架以解决无任务边界的持续学习问题 large language model
19 A Single Suffix to Break Them All: Basin-Aware Jailbreaks for Merged Model Families 提出Basin-Aware Jailbreak以解决模型合并安全隐患问题 foundation model
20 Leveraging a Foundation Model for the EEG-Based Diagnosis of Alzheimer's Disease 利用基础模型解决阿尔茨海默病的EEG诊断问题 foundation model
21 Towards Large-Scale Heterogeneous Data Organization for Scientific Foundation Models: A Nuclear Fusion Case Study 提出大规模异构数据组织方法以解决核聚变领域挑战 foundation model
22 The Calls are Coming from Inside the Model: Investigating Probe-based Detection of Tool-Calling Errors in LLMs 提出探测工具调用错误的新方法以提升LLM的可靠性 large language model

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
23 Shared Actors Need Not Share Critics: Effects of Value Mismatch in Parallel Reinforcement Learning 提出环境索引以解决并行强化学习中的价值不匹配问题 bipedal biped reinforcement learning
24 Arrive and Survive: Scaling Safe Goal-Conditioned Policy Learning from One-Bit Failure Signals 提出安全对比强化学习以解决失败终止下的目标导向策略学习问题 locomotion reinforcement learning policy learning ✅
25 Chart2SVG: Editable SVG Generation from Raster Chart Images 提出Chart2SVG以解决静态图表转换为可编辑SVG的问题 manipulation large language model multimodal

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#题目一句话要点标签🔗⭐
26 SecureDrive-FL: Joint Differential Privacy and Gradient-Aware Selective Homomorphic Encryption for Federated Driver Monitoring 提出SecureDrive-FL解决联邦学习中的隐私与安全问题 OMOMO

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

#题目一句话要点标签🔗⭐
27 Self-Augmented Diffusion Guidance for Physics-Informed Generation 提出自增强扩散引导以解决物理约束生成问题 spatiotemporal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页