cs.LG（2025-12-25）
📊 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) (5)
支柱九：具身大模型 (Embodied Foundation Models) (4 🔗1)
支柱四：生成式动作 (Generative Motion) (1)
支柱八：物理动画 (Physics-based Animation) (1)
🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)
| # | 题目 | 一句话要点 | 标签 | 🔗 | ⭐ |
|---|---|---|---|---|---|
| 1 | Rethinking Output Alignment For 1-bit Post-Training Quantization of Large Language Models | 提出一种面向大语言模型1比特后训练量化的输出对齐方法，提升量化性能。 | distillation large language model | ||
| 2 | Horizon Reduction as Information Loss in Offline Reinforcement Learning | 揭示离线强化学习中Horizon Reduction导致信息损失的根本原因与结构性失效模式 | reinforcement learning offline RL offline reinforcement learning | ||
| 3 | Videos are Sample-Efficient Supervisions: Behavior Cloning from Videos via Latent Representations | 提出BCV-LR框架，通过视频中的潜在表征实现高效的行为克隆，解决交互样本不足的问题。 | reinforcement learning policy learning imitation learning | ||
| 4 | Global-Graph Guided and Local-Graph Weighted Contrastive Learning for Unified Clustering on Incomplete and Noise Multi-View Data | 提出全局图引导和局部图加权对比学习框架，解决不完整和噪声多视图聚类问题 | contrastive learning | ||
| 5 | AVP-Fusion: Adaptive Multi-Modal Fusion and Contrastive Learning for Two-Stage Antiviral Peptide Identification | AVP-Fusion：融合自适应多模态和对比学习的两阶段抗病毒肽识别方法 | contrastive learning |
🔬 支柱九：具身大模型 (Embodied Foundation Models) (4 篇)
| # | 题目 | 一句话要点 | 标签 | 🔗 | ⭐ |
|---|---|---|---|---|---|
| 6 | Physic-HM: Restoring Physical Generative Logic in Multimodal Anomaly Detection via Hierarchical Modulation | Physic-HM：通过层级调制恢复物理生成逻辑的多模态异常检测 | multimodal | ||
| 7 | RefineBridge: Generative Bridge Models Improve Financial Forecasting by Foundation Models | RefineBridge：生成桥模型通过基础模型改进金融预测 | foundation model | ||
| 8 | nncase: An End-to-End Compiler for Efficient LLM Deployment on Heterogeneous Storage Architectures | nncase：用于异构存储架构上高效LLM部署的端到端编译器 | large language model | ✅ | |
| 9 | Hierarchical Stacking Optimization Using Dirichlet's Process (SoDip): Towards Accelerated Design for Graft Polymerization | 提出SoDip框架以解决辐射诱导接枝聚合的可重复性问题 | multimodal |
🔬 支柱四：生成式动作 (Generative Motion) (1 篇)
| # | 题目 | 一句话要点 | 标签 | 🔗 | ⭐ |
|---|---|---|---|---|---|
| 10 | Co-GRPO: Co-Optimized Group Relative Policy Optimization for Masked Diffusion Model | 提出Co-GRPO，协同优化Masked Diffusion Model及其解码策略，提升生成质量。 | MDM |
🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)
| # | 题目 | 一句话要点 | 标签 | 🔗 | ⭐ |
|---|---|---|---|---|---|
| 11 | RIPCN: A Road Impedance Principal Component Network for Probabilistic Traffic Flow Forecasting | RIPCN：一种用于概率交通流预测的道路阻抗主成分网络 | spatiotemporal |