cs.LG（2025-12-13）

📊 共 18 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (9 🔗3) 支柱九：具身大模型 (Embodied Foundation Models) (7) 支柱八：物理动画 (Physics-based Animation) (1) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱二：RL算法与架构 (RL & Architecture) (9 篇)

#题目一句话要点标签🔗⭐
1 RAST-MoE-RL: A Regime-Aware Spatio-Temporal MoE Framework for Deep Reinforcement Learning in Ride-Hailing 提出RAST-MoE-RL框架，解决网约车中复杂时空动态下的自适应延迟匹配问题。 reinforcement learning deep reinforcement learning spatiotemporal
2 Knowledge-Guided Masked Autoencoder with Linear Spectral Mixing and Spectral-Angle-Aware Reconstruction 提出知识引导的掩码自编码器，融合线性光谱混合模型与光谱角感知重建，提升表征学习。 masked autoencoder geometric consistency
3 Anchoring Values in Temporal and Group Dimensions for Flow Matching Model Alignment 提出VGPO，通过时序和群体维度锚定价值，提升Flow Matching模型对齐效果。 flow matching large language model ✅
4 EEG-DLite: Dataset Distillation for Efficient Large EEG Model Training EEG-DLite：脑电数据蒸馏框架，高效训练大规模脑电模型 distillation foundation model ✅
5 MolGuidance: Advanced Guidance Strategies for Conditional Molecular Generation with Flow Matching MolGuidance：利用Flow Matching进行条件分子生成的高级引导策略 flow matching classifier-free guidance
6 Cross-Modal Representational Knowledge Distillation for Enhanced Spike-Informed LFP Modeling 提出跨模态表征知识蒸馏方法，提升基于Spike信息的LFP建模性能 distillation spatiotemporal
7 DeepVekua: Geometric-Spectral Representation Learning for Physics-Informed Fields DeepVekua：结合几何深度学习与谱分析求解稀疏数据下的偏微分方程 representation learning implicit representation ✅
8 TS-DP: Reinforcement Speculative Decoding For Temporal Adaptive Diffusion Policy Acceleration 提出TS-DP，通过强化学习推测解码加速时序自适应扩散策略，实现实时具身控制。 diffusion policy
9 ElasticVR: Elastic Task Computing in Multi-User Multi-Connectivity Wireless Virtual Reality (VR) Systems ElasticVR：面向多用户多连接无线VR系统的弹性任务计算框架 reinforcement learning deep reinforcement learning

🔬 支柱九：具身大模型 (Embodied Foundation Models) (7 篇)

#题目一句话要点标签🔗⭐
10 The Data Efficiency Frontier of Financial Foundation Models: Scaling Laws from Continued Pretraining 通过持续预训练提升金融领域大语言模型的数据效率，揭示模型规模扩展规律。 large language model foundation model
11 On Harnessing Idle Compute at the Edge for Foundation Model Training Cleave：利用边缘空闲算力高效训练大模型，媲美云端性能 foundation model
12 Can Graphs Improve Tabular Foundation Models? BOLERO：利用图结构先验提升预训练表格数据Transformer模型性能 foundation model
13 BOOST: BOttleneck-Optimized Scalable Training Framework for Low-Rank Large Language Models BOOST：面向低秩大语言模型的瓶颈优化可扩展训练框架 large language model
14 Dynamical modeling of nonlinear latent factors in multiscale neural activity with real-time inference 提出一种多尺度神经活动非线性动态建模框架，用于实时推断目标变量。 multimodal
15 Keep the Lights On, Keep the Lengths in Check: Plug-In Adversarial Detection for Time-Series LLMs in Energy Forecasting 提出即插即用的对抗样本检测框架，保障能源预测中时间序列LLM的安全性。 large language model
16 Citation-Grounded Code Comprehension: Preventing LLM Hallucination Through Hybrid Retrieval and Graph-Augmented Context 提出混合检索与图增强上下文的代码理解方法，解决LLM代码幻觉问题。 large language model

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

#题目一句话要点标签🔗⭐
17 BaRISTA: Brain Scale Informed Spatiotemporal Representation of Human Intracranial Neural Activity BaRISTA：提出脑尺度感知的时空Transformer模型，提升人脑颅内神经活动解码性能。 spatiotemporal foundation model

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
18 Composite Classifier-Free Guidance for Multi-Modal Conditioning in Wind Dynamics Super-Resolution 提出复合无分类器引导（CCFG）方法，用于提升风场超分辨率重建中多模态条件扩散模型的性能。 classifier-free guidance

⬅️ 返回 cs.LG 首页 · 🏠 返回主页