cs.LG（2026-10-26）

📊 共 31 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (19) 支柱九：具身大模型 (Embodied Foundation Models) (10 🔗2) 支柱一：机器人控制 (Robot Control) (2)

🔬 支柱二：RL算法与架构 (RL & Architecture) (19 篇)

#题目一句话要点标签🔗⭐
1 TimeWorld: Towards Action-Conditioned Multimodal World Models for Time Series Forecasting 提出TimeWorld以解决时间序列预测中的行动依赖建模问题 world model world models contrastive learning
2 Beyond Modeling Order: Dual-Path Spatiotemporal Representation Learning for Cross-Subject EEG Emotion Recognition 提出双路径时空对比学习以解决跨主体EEG情感识别问题 representation learning contrastive learning motion representation
3 When Fine-Grained Credit Assignment Helps and Hurts: Multimodal Agentic Reinforcement Learning through Video Editing 提出多模态强化学习方法以优化视频编辑任务中的信用分配问题 reinforcement learning distillation privileged information
4 Video World Models Need Predictable Latents 提出可预测潜在空间以解决视频世界模型的长期预测问题 world model world models representation learning
5 AgriDefend: A $205 Offline Edge Device that Detects Crop Stress Before Visible Symptoms via Multimodal AI Fusion 提出AgriDefend以解决作物应激早期检测问题 MAE multimodal
6 Reinforcement Learning Done Right for Explicit Confidence Targets 提出RLEC以解决大语言模型的信心目标控制问题 reinforcement learning large language model
7 End-to-End MXFP8 Reinforcement Learning on Blackwell: What Matching Buys, When It Pays Off, and How It Fails 提出MXFP8强化学习以优化大语言模型训练效率 reinforcement learning large language model
8 Multi-Objective Reinforcement Learning Using Routed Ensembles and Trajectory Attribution 提出基于路由集成和轨迹归因的多目标强化学习方法 reinforcement learning policy learning
9 TransECG-SSL: Hybrid Self-Supervised Representation Learning for ECG Arrhythmia Classification 提出TransECG-SSL以解决ECG心律失常分类的跨数据集泛化问题 representation learning contrastive learning
10 AutoPRL: Automated Path Representation Learning via a Learnable View Generator 提出AutoPRL以解决路径表示学习中的视图生成和训练效率问题 representation learning contrastive learning
11 Transporting Quantiles to Codebook: Scalable Vector Quantization without Codebook Collapse 提出FlowVQ以解决向量量化中的代码本崩溃问题 flow matching representation learning VQ-VAE
12 CounterFlow: Target-Conditioned Flow Matching for Visual Counterfactuals 提出CounterFlow以生成视觉反事实解释，提升模型可解释性 flow matching
13 Preference-Guided Cooperation in Multi-Objective Multi-Agent Reinforcement Learning 提出MIRL以解决多目标多智能体强化学习中的合作问题 reinforcement learning
14 MiRL: Error-Targeted Reinforcement Learning for Correcting Terminal Agent Misbehaviors 提出MiRL以解决终端代理错误行为问题 reinforcement learning
15 S3AC: Stabilized, Selective, and Scalable Actor-Critic for LLM Reasoning 提出S3AC以解决长链推理中的批评者训练问题 reinforcement learning chain-of-thought
16 GPDPO: Learning from Reward Conflicts via Group Pareto-Dominance Policy Optimization 提出GPDPO以解决多奖励信号冲突问题 reinforcement learning large language model
17 Multi-Horizon Future Discovery for Physics-Informed Exploration 提出HTC²框架以增强物理控制环境中的探索能力 PPO SAC representation learning
18 Reinforcement learning for constrained stochastic linear--quadratic control in continuous time 提出强化学习方法解决约束随机线性-二次控制问题 reinforcement learning
19 Stable Continuous-Time Consistency Distillation: An Empirical Study with a Multistep Extension 提出稳定的连续时间一致性蒸馏方法以解决生成模型训练不稳定问题 distillation

🔬 支柱九：具身大模型 (Embodied Foundation Models) (10 篇)

#题目一句话要点标签🔗⭐
20 OneEEG: A Unified Signal-Visual Multimodal Language Model with Task-Aware Mixture of Projectors for Multi-Task EEG Understanding 提出OneEEG以解决多任务EEG理解的碎片化问题 large language model multimodal
21 TUON: A Tensor-Aware Optimizer Utilizing Tensor-SVD and Polarization for Large Language Model Pretraining 提出TUON优化器以解决大规模语言模型预训练效率问题 large language model
22 Auto-Research Arena: Benchmarking and Advancing Auto-Research in Language Modeling 提出Auto-Research Arena以评估和推动语言建模中的自主研究 foundation model ✅
23 Work Agents: Systems, Capabilities, and Optimization 提出工作代理系统框架以优化数字工作执行 foundation model
24 Selective Tie Learning: Mitigating Spurious Correlations in Preference Optimization 提出选择性平衡学习以缓解偏好优化中的虚假相关性问题 large language model
25 A Hybrid Framework for Explainable Aviation Risk: Causal Chain Extraction from Safety Reports with Physics-Based Failure Quantification and Interventional Analysis 提出混合框架以实现航空风险的可解释性分析 large language model
26 Prior Predictive Matching for Bayesian LoRA 提出闭式先验预测估计器以解决LoRA超参数选择问题 foundation model
27 BellowsKV: One Latent for Many Layers, Expanded on Demand 提出BellowsKV以解决KV缓存压缩精度不足问题 large language model ✅
28 When Respondents Are Scarce: LLM-Derived Semantic Priors for Bayesian Confirmatory Factor Analysis 提出语义经验更新框架以解决小样本贝叶斯确认性因子分析问题 large language model
29 Mechanistic Insights into LoRA: Layer Sparsity for Adaptive Fine-Tuning via Path Patching 提出Adaptive-LoRA以解决大语言模型自适应微调的效率问题 large language model

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
30 Adapt for Planning, Not just Prediction: Optimistic Test-Time World Model Adaptation 提出乐观测试时世界模型适应以解决规划精度下降问题 manipulation world model world models
31 High-resolution tactile observation makes hidden racket-control mechanisms learnable 提出高分辨率触觉观察以揭示羽毛球拍控制机制 manipulation spatiotemporal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页