cs.LG（2026-05-25）
📊 共 4 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) (2 篇)
| # | 题目 | 一句话要点 | 标签 | 🔗 | ⭐ |
|---|---|---|---|---|---|
| 1 | UWM-JEPA: Predictive World Models That Imagine in Belief Space | 提出UWM-JEPA，利用密度矩阵和酉预测器提升部分可观测环境下的世界模型预测能力 | world model world models JEPA | ||
| 2 | When Self-Belief Misleads: Active Label Acquisition for Reinforcement Learning with Verifiable Rewards | 提出RLAVR框架，通过主动学习策略提升可验证奖励强化学习的性能与稳定性。 | reinforcement learning large language model | ✅ |
🔬 支柱九：具身大模型 (Embodied Foundation Models) (2 篇)
| # | 题目 | 一句话要点 | 标签 | 🔗 | ⭐ |
|---|---|---|---|---|---|
| 3 | Prism: A Plug-in Reproducible Infrastructure for Scalable Multimodal Continual Instruction Tuning | Prism：用于可扩展多模态持续指令微调的插件式可复现基础设施 | large language model multimodal instruction following | ✅ | |
| 4 | RotMoLE: Enhancing Mixture of Low-Rank Experts through Rotational Gating Mechanism | RotMoLE：通过旋转门机制增强低秩专家混合模型，提升复杂场景下的知识学习能力 | large language model |