cs.LG（2026-02-07）

📊 共 19 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (10) 支柱二：RL算法与架构 (RL & Architecture) (5) 支柱一：机器人控制 (Robot Control) (2 🔗1) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱八：物理动画 (Physics-based Animation) (1)

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

#题目一句话要点标签🔗⭐
1 Controllable Value Alignment in Large Language Models through Neuron-Level Editing NeVA：通过神经元级编辑实现大语言模型中可控的价值观对齐 large language model
2 Towards Robust Scaling Laws for Optimizers 提出优化器鲁棒缩放律，实现不同优化器在LLM预训练中的公平比较。 large language model
3 Rational Transductors 提出Rational Transductors，解决Transformer在序列逻辑和状态追踪上的泛化难题。 chain-of-thought
4 Astro: Activation-guided Structured Regularization for Outlier-Robust LLM Post-Training Quantization 提出Astro框架以解决LLM后训练量化中的异常值问题 large language model
5 Gaussian Match-and-Copy: A Minimalist Benchmark for Studying Transformer Induction 提出高斯匹配复制基准测试，用于研究Transformer的归纳能力 large language model
6 Hyperparameter Transfer Laws for Non-Recurrent Multi-Path Neural Networks 提出基于图的有效深度概念，实现非循环多路径神经网络超参数的零样本迁移。 zero-shot transfer
7 On the Importance of a Multi-Scale Calibration for Quantization 提出MaCa：一种多尺度校准方法，提升LLM量化在变长输入下的精度。 large language model
8 Sign-Based Optimizers Are Effective Under Heavy-Tailed Noise SignSGD优化器在重尾噪声下表现优异，为大模型训练提供理论支撑 large language model
9 Revisiting Robustness for LLM Safety Alignment via Selective Geometry Control ShaPO：通过选择性几何控制提升LLM安全对齐的鲁棒性 large language model
10 Parallel Track Transformers: Enabling Fast GPU Inference with Reduced Synchronization 提出并行轨道Transformer，减少同步操作，加速GPU推理。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
11 Proximal Action Replacement for Behavior Cloning Actor-Critic in Offline Reinforcement Learning 提出近端动作替换(PAR)方法，突破离线强化学习中行为克隆的性能上限 reinforcement learning TD3 offline RL
12 CoMI-IRL: Contrastive Multi-Intention Inverse Reinforcement Learning 提出CoMI-IRL，一种无需先验知识的多意图逆强化学习框架 reinforcement learning inverse reinforcement learning
13 Do We Need Adam? Surprisingly Strong and Sparse Reinforcement Learning with SGD in LLMs 在LLM的RL训练中，SGD表现优于AdamW，并实现极高的参数稀疏性 reinforcement learning large language model
14 Efficient Planning in Reinforcement Learning via Model Introspection 提出基于模型内省的强化学习高效规划方法，连接RL与经典规划 reinforcement learning
15 AI-Driven Predictive Modelling for Groundwater Salinization in Israel 利用AI预测以色列地下水盐碱化，揭示关键驱动因素并降低模型不确定性。 predictive model

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
16 Scalable Dexterous Robot Learning with AR-based Remote Human-Robot Interactions 提出基于AR远程人机交互的可扩展灵巧机器人学习框架，提升操作任务效率。 manipulation reinforcement learning behavior cloning ✅
17 Data-Aware and Scalable Sensitivity Analysis for Decision Tree Ensembles 提出数据感知的决策树集成敏感性分析框架，提升模型可靠性和公平性。 manipulation

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
18 Dense Feature Learning via Linear Structure Preservation in Medical Data 提出密集特征学习框架，通过线性结构保持提升医疗数据表征的泛化性和稳定性。 structure preservation multimodal

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

#题目一句话要点标签🔗⭐
19 Learned Finite Element-based Regularization of the Inverse Problem in Electrocardiographic Imaging 提出基于学习的有限元正则化方法，提升心电成像逆问题的重建精度。 spatiotemporal

⬅️ 返回 cs.LG 首页 · 🏠 返回主页