cs.LG（2026-09-18）

📊 共 20 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (9) 支柱九：具身大模型 (Embodied Foundation Models) (9 🔗1) 支柱一：机器人控制 (Robot Control) (2 🔗1)

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

#题目一句话要点标签🔗⭐
1 Beyond Kinematics: Benchmarking Simulation Fidelity for Muscle-Driven Imitation Learning 比较两种运动模仿强化学习管道以提升生理现实性 reinforcement learning policy learning imitation learning
2 OneBid: A Unified Auto-Bidding Foundation Model for Diverse oCPX Advertising Scenarios 提出OneBid以解决oCPX广告场景的统一自动竞价问题 reinforcement learning decision transformer foundation model
3 Intervention Granularity Matters: Coherent Treatment Bundles in Counterfactual Simulation with Clinical World Models 提出干预粒度的重要性以优化反事实模拟 world model world models JEPA
4 Deep Reinforcement Learning with Buffered Quantile Objectives 提出深度强化学习框架以解决量化目标优化问题 reinforcement learning deep reinforcement learning PPO
5 Beyond Gaussian Worlds: Latent Geometry Matters for JEPAs 扩展JEPAs以解决潜在几何对表示恢复的影响 JEPA Joint-Embedding Predictive Architecture joint-embedding predictive architecture
6 $λ$-Controlled GRPO: Turning Flow-Matching Ratio Instability into a Budgeted Resource 提出$λ$-Controlled GRPO以解决流匹配比率不稳定问题 reinforcement learning flow matching
7 On Repulsive and Attractive Teachers: Separating Correctness from Behavior in Self-Distillation 提出对比自蒸馏方法以解决行为偏差问题 distillation privileged information
8 Time series generation with spectrally aligned latent flow matching 提出光谱对齐的潜在流匹配模型以解决时间序列生成问题 flow matching
9 IncentRL: The Trade-Off Between Preference Guidance and Task Performance 提出IncentRL框架以解决偏好引导与任务性能之间的权衡问题 reinforcement learning reward shaping

🔬 支柱九：具身大模型 (Embodied Foundation Models) (9 篇)

#题目一句话要点标签🔗⭐
10 M2G-LLM: Enhancing Clinical Prediction via Multimodal Graph Reasoning and LLM Context Injection 提出M2G-LLM以解决临床预测中的多模态数据整合问题 large language model multimodal
11 Detecting Pretraining Data in Large Language Models from a Free-Energy Perspective 提出基于自由能视角的预训练数据检测方法以解决现有检测不足问题 large language model
12 Watermarkable Multi-Draft Speculative Sampling via Poisson Processes 提出基于泊松过程的多草稿推测采样算法以解决效率与水印问题 large language model
13 GraphSkillEvo: Evolutionary Optimization of Graph-Structured Agent Skills 提出GraphSkillEvo以解决图结构代理技能优化问题 large language model ✅
14 SpecQuant: Speculative Decoding with Multi-Parent Quantization for Adaptive LLM Inference 提出SpecQuant以解决大语言模型推理效率问题 large language model
15 Understanding LLM Quantization through Activation-Guided Compensation and Orthogonal Residuals 通过激活引导补偿和正交残差提出LLM量化新方法 large language model
16 Probabilistic Forecasting of Business Process Executions with Neural Temporal Point Processes 提出基于神经时间点过程的业务流程执行概率预测方法 TAMP
17 Knowledge-Graph-Augmented Chronos-2 for HEC-RAS Surrogate Forecasting 提出KG-Chronos-2以提升HEC-RAS水面高度预测精度 foundation model
18 TierKV: Long-Context On-Device LLMs via Predictive Multi-Tier KV Caching 提出TierKV以解决移动设备长上下文推理的内存瓶颈问题 large language model

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
19 Benchmarking World Models for Continual Learning on Compositional Tasks 提出组合持续学习基准以解决世界模型知识重用问题 manipulation world model world models ✅
20 GEM-MPC: Balancing Exploration and Exploitation through Expert-Guided Planning 提出GEM-MPC以解决高维控制中的探索与利用平衡问题 MPC reinforcement learning distillation

⬅️ 返回 cs.LG 首页 · 🏠 返回主页