cs.LG（2026-10-20）

📊 共 32 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (15 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (14 🔗1) 支柱八：物理动画 (Physics-based Animation) (2) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (15 篇)

#题目一句话要点标签🔗⭐
1 Orthogonal Boundary Unlearning for Dual-Use Concepts in Large Language Models 提出正交边界去学习以解决双重用途概念问题 large language model foundation model
2 Training-Free Continual Learning for Multimodal Forecasting via Prompt Evolution 提出无训练持续学习方法以解决多模态预测问题 foundation model multimodal
3 Turning TS2Vec into a Foundation Model for Time Series Classification 将TS2Vec转变为时间序列分类的基础模型 foundation model
4 Parameter-Efficient Adaptation of Time Series Foundation Models for Electrical Load Forecasting 提出低秩适应方法以提高电力负荷预测的效率与准确性 foundation model
5 BEACON: Bridging Multimodal Semantics and Asynchronous Propagation for Cloud System Anomaly Detection 提出BEACON框架以解决云系统异常检测中的多模态语义与异步传播问题 multimodal
6 Efficient Edge-aware Attention Network for Graph Generation 提出边缘感知注意力网络以解决图生成中的可扩展性问题 foundation model
7 The Token-Energy Law of LLM Decode: Active Bytes and Realized Time Account for Joules per Token 提出Token-Energy法则以优化大语言模型的能效问题 large language model
8 When Do LLMs Help with Feature Selection? A Systematic Empirical Study 探讨大型语言模型在特征选择中的有效性与局限性 large language model
9 Rewards for Budget-Aware Multi-Turn Agents That Adapt to Arbitrary Test-Time Constraints 提出预算感知的多轮智能体以适应任意测试时间约束 large language model
10 Don't Skip the Baseline: A Controlled Comparison of Dynamic Expert Skipping in MoEs 提出动态专家跳过方法的基线比较以优化MoE模型 multimodal
11 MuPACE: Learning to Pace Muon for LoRA Fine-Tuning 提出MuPACE以优化LoRA微调中的学习率选择问题 large language model
12 Poisoning-Based Backdoors Are Not Cryptographic 揭示基于毒化的后门攻击无法实现密码学不对称性 large language model
13 SliMOO: Interpretable Multi-Objective Evolutionary Search for LLM Depth Pruning 提出SliMOO以解决大语言模型深度剪枝效率问题 large language model ✅
14 Foundation-Style Pretraining for Scalable and Transferable Distributed PV Forecasting 提出基础风格预训练框架以解决分布式光伏预测的可扩展性问题 foundation model
15 Two Is Enough: Binary Classification Priors Suffice for Multi-Class In-Context Tabular Prediction 提出二分类先验以提升多分类表格预测性能 foundation model

🔬 支柱二：RL算法与架构 (RL & Architecture) (14 篇)

#题目一句话要点标签🔗⭐
16 Bayesian Neural Networks via Dirichlet Process Priors on the Data-Generating Distribution for Randomized Value Functions in Deep Reinforcement Learning 提出基于Dirichlet过程先验的贝叶斯神经网络以解决强化学习中的不确定性问题 reinforcement learning deep reinforcement learning
17 Projected Distillation: On Learning to Solve Hard Problems from Close, Correct Targets 提出投影蒸馏以解决困难问题的学习挑战 policy learning behavior cloning distillation
18 TPO: Tri-level Distributionally Robust Learning for OOD Direct Preference Optimization 提出TPO框架以解决OOD偏好优化中的分布不匹配问题 DPO direct preference optimization
19 Reframing Object-Centric Reinforcement Learning: From Absolute to Relative Representations 提出ORCA与ORBiT以解决对象中心强化学习中的绝对表示问题 reinforcement learning PPO
20 Ask to Learn: Query-Based Collaborative Reinforcement Learning 提出查询驱动的协作强化学习以克服离线学习局限 reinforcement learning offline reinforcement learning
21 MiCLA: Mixture-to-Constituent Latent Alignment for Polyphonic Audio Representation Learning 提出MiCLA以解决多音频表示学习中的混合-成分关系建模问题 representation learning distillation
22 LocaLeJEPA: Maximum Entropy Representations under Geometric Priors 提出LocaLeJEPA以解决JEPA方法中的表示崩溃问题 JEPA Joint-Embedding Predictive Architecture joint-embedding predictive architecture ✅
23 GraphCast-Mamba: Learning Non-Markovian Corrections for Weather Forecasting 提出GraphCast-Mamba以解决天气预报中的非马尔可夫性问题 Mamba
24 Model- and Layer-Level Closed-Loop Forecasting with SSM-Based Spiking Neural Networks 提出闭环预测策略以解决SNN时间序列预测问题 SSM
25 Adaptive Pullback: Reweighting Supervision at Penalty Jumps in On-Policy Distillation 提出自适应回退方法解决在线蒸馏中的惩罚跳跃问题 distillation
26 Causal Discovery for Linear Models with Latent Confounding via Multi-domain Identifiable Representation Learning 提出LinADMG框架以解决潜在混淆的线性因果发现问题 representation learning
27 The Shape of the DPO Hyperparameter Landscape: What to Measure, How It Moves with Scale, and When Not to Transfer It 研究DPO超参数景观以优化语言模型后训练 DPO
28 Self-Evolving Agents through Privileged Retrospective Distillation 提出自我演化代理以解决缺乏真实奖励的问题 distillation
29 From Outcomes to Steps: Reliable Success-Aware Critic Learning for Long-Horizon Language Agents 提出SCALE框架以解决长时间语言代理的成功预测问题 reinforcement learning PPO

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗⭐
30 PULSE: Personalized Opportunistic Learning for Wearable Devices 提出PULSE框架以解决可穿戴设备个性化学习问题 PULSE
31 Geometry-Agnostic Data-Driven Discovery of Extreme Scenarios 提出一种几何无关的数据驱动极端场景发现方法 spatiotemporal PULSE

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
32 From POMDP Theory to Deep RL with Particle Filters 提出基于粒子滤波的深度强化学习算法以解决部分可观测环境中的决策问题 MPC reinforcement learning latent dynamics

⬅️ 返回 cs.LG 首页 · 🏠 返回主页