cs.LG（2025-06-28）

📊 共 11 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) (3) 支柱八：物理动画 (Physics-based Animation) (3) 支柱一：机器人控制 (Robot Control) (2 🔗1) 支柱九：具身大模型 (Embodied Foundation Models) (2) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#	题目	一句话要点	标签	🔗	⭐
1	Infinite Sampling: Efficient and Stable Grouped RL Training for Large Language Models	提出Infinite Sampling框架，解决LLM分组强化学习训练中内存瓶颈问题。	reinforcement learning large language model
2	FairMarket-RL: LLM-Guided Fairness Shaping for Multi-Agent Reinforcement Learning in Peer-to-Peer Markets	FairMarket-RL：基于LLM引导的强化学习框架，用于点对点市场中的公平性塑造	reinforcement learning reward shaping large language model
3	Fragile, Robust, and Antifragile: A Perspective from Parameter Responses in Reinforcement Learning Under Stress	提出基于参数响应的强化学习鲁棒性分析框架，提升策略抗干扰能力	reinforcement learning PPO

🔬 支柱八：物理动画 (Physics-based Animation) (3 篇)

#	题目	一句话要点	标签	🔗	⭐
4	On Universality of Non-Separable Approximate Message Passing Algorithms	研究非可分近似消息传递算法的普适性，扩展其在非高斯数据下的适用范围	AMP
5	Interpretable Time Series Autoregression for Periodicity Quantification	提出稀疏自回归模型以量化时间序列的周期性	spatiotemporal
6	Robust Tensor Completion via Gradient Tensor Nulclear L1-L2 Norm for Traffic Data Recovery	提出梯度张量核L1-L2范数以解决交通数据恢复问题	spatiotemporal

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
7	Beyond Parallelism: Synergistic Computational Graph Effects in Multi-Head Attention	揭示多头注意力机制中计算图协同效应，超越并行计算的优势	manipulation large language model	✅
8	Towards Time Series Generation Conditioned on Unstructured Natural Language	提出一种基于扩散模型和语言模型的自然语言条件时间序列生成方法	manipulation

🔬 支柱九：具身大模型 (Embodied Foundation Models) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
9	Spectra 1.1: Scaling Laws and Efficient Inference for Ternary Language Models	Spectra 1.1：通过三元语言模型和高效推理加速大规模语言模型部署	large language model
10	BEST-Route: Adaptive LLM Routing with Test-Time Optimal Compute	BEST-Route：基于测试时最优计算的自适应LLM路由框架	large language model

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#	题目	一句话要点	标签	🔗	⭐
11	Multimodal Atmospheric Super-Resolution With Deep Generative Models	利用深度生成模型实现多模态大气超分辨率重建	sparse sensors spatiotemporal multimodal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页