cs.LG（2026-03-16）

📊 共 24 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) (14) 支柱九：具身大模型 (Embodied Foundation Models) (9 🔗2) 支柱一：机器人控制 (Robot Control) (1)

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

#题目一句话要点标签🔗⭐
1 Towards Foundation Models for Consensus Rank Aggregation 提出基于Transformer的Kemeny Transformer，用于高效共识排序聚合。 reinforcement learning foundation model
2 Deep Reinforcement Learning for Fano Hypersurfaces 提出深度强化学习算法以探索Fano超曲面 reinforcement learning deep reinforcement learning
3 Effective Distillation to Hybrid xLSTM Architectures 提出一种有效的蒸馏流程，将大型语言模型提炼到混合xLSTM架构，实现性能匹配甚至超越。 distillation large language model
4 Mamba-3: Improved Sequence Modeling using State Space Principles Mamba-3：利用状态空间模型原理改进序列建模，提升推理效率与模型质量。 Mamba SSM state space model
5 Amplification Effects in Test-Time Reinforcement Learning: Safety and Reasoning Vulnerabilities 研究表明，测试时强化学习易受有害提示注入攻击，导致安全性放大和推理能力下降。 reinforcement learning large language model
6 Safe Flow Q-Learning: Offline Safe Reinforcement Learning with Reachability-Based Flow Policies 提出SafeFQL，结合可达性分析与流策略，解决离线安全强化学习问题 reinforcement learning offline RL
7 TabKD: Tabular Knowledge Distillation through Interaction Diversity of Learned Feature Bins TabKD：通过学习特征箱交互多样性的表格知识蒸馏 distillation
8 TrajFlow: Nation-wide Pseudo GPS Trajectory Generation with Flow Matching Models TrajFlow：基于Flow Matching模型的全国范围伪GPS轨迹生成 flow matching
9 Photonic Quantum-Enhanced Knowledge Distillation 提出光子量子增强知识蒸馏(PQKD)框架，利用光子电路提升模型压缩性能。 distillation
10 Sample-Efficient Hypergradient Estimation for Decentralized Bi-Level Reinforcement Learning 提出基于Boltzmann协方差技巧的超梯度估计方法，解决去中心化双层强化学习问题。 reinforcement learning
11 Dataset Distillation Efficiently Encodes Low-Dimensional Representations from Gradient-Based Learning of Non-Linear Tasks 理论分析数据集蒸馏，揭示梯度学习非线性任务的低维表征高效编码 distillation
12 DeFRiS: Silo-Cooperative IoT Applications Scheduling via Decentralized Federated Reinforcement Learning DeFRiS：通过去中心化联邦强化学习实现Silo协同物联网应用调度 reinforcement learning
13 Faster Inference of Flow-Based Generative Models via Improved Data-Noise Coupling 提出LOOM-CFM，通过跨Minibatch优化数据-噪声耦合加速Flow模型推理。 flow matching distillation
14 Sampling-guided exploration of active feature selection policies 提出基于采样指导的主动特征选择策略，提升高维数据分类性能并降低特征获取成本。 reinforcement learning predictive model

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

#题目一句话要点标签🔗⭐
15 CAMD: Coverage-Aware Multimodal Decoding for Efficient Reasoning of Multimodal Large Language Models 提出CAMD：一种覆盖感知的多模态解码方法，用于提升多模态大语言模型的推理效率 large language model multimodal
16 IntegratingWeather Foundation Model and Satellite to Enable Fine-Grained Solar Irradiance Forecasting 提出Baguan-solar，融合气象大模型与卫星数据，实现精细化太阳辐照度预测 foundation model multimodal ✅
17 Physics-informed fine-tuning of foundation models for partial differential equations 提出物理信息微调框架，高效利用偏微分方程基础模型解决数据稀缺问题。 foundation model
18 Multimodal Deep Learning for Early Prediction of Patient Deterioration in the ICU: Integrating Time-Series EHR Data with Clinical Notes 提出多模态深度学习方法，融合时间序列EHR数据与临床笔记，用于ICU患者病情恶化早期预测。 multimodal
19 HorizonMath: Measuring AI Progress Toward Mathematical Discovery with Automatic Verification HorizonMath：利用自动验证评估AI在数学发现方面的进展 large language model
20 Not All Invariants Are Equal: Curating Training Data to Accelerate Program Verification with SLMs 提出Wonda数据管线，通过高质量数据训练SLM加速程序验证中的循环不变式生成。 large language model
21 Establishing Construct Validity in LLM Capability Benchmarks Requires Nomological Networks 提出名义网络以解决LLM能力基准的构念有效性问题 large language model
22 Cross-RAG: Zero-Shot Retrieval-Augmented Time Series Forecasting via Cross-Attention 提出Cross-RAG，通过交叉注意力实现零样本检索增强时间序列预测 foundation model ✅
23 Design Space of Self--Consistent Electrostatic Machine Learning Interatomic Potentials 提出自洽静电机器学习原子间势的设计空间，解决长程静电效应建模难题。 foundation model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
24 Orthogonal Subspace Clustering: Enhancing High-Dimensional Data Analysis through Adaptive Dimensionality Reduction and Efficient Clustering 提出正交子空间聚类以解决高维数据分析问题 OSC

⬅️ 返回 cs.LG 首页 · 🏠 返回主页