cs.LG（2025-12-11）

📊 共 9 篇论文

🎯 兴趣领域导航

#	题目	一句话要点	标签
1	Digital Twin Supervised Reinforcement Learning Framework for Autonomous Underwater Navigation	提出基于数字孪生监督强化学习的水下自主导航框架	reinforcement learning deep reinforcement learning PPO
2	Refining Graphical Neural Network Predictions Using Flow Matching for Optimal Power Flow with Constraint-Satisfaction Guarantee	提出基于流匹配的图神经网络优化方法，保障约束条件下的最优潮流计算	flow matching penetration
3	Bandwidth-constrained Variational Message Encoding for Cooperative Multi-agent Reinforcement Learning	提出BVME：带宽约束下多智能体强化学习的变分消息编码方法	reinforcement learning
4	Adaptive Replay Buffer for Offline-to-Online Reinforcement Learning	提出自适应回放缓存ARB，解决离线到在线强化学习的数据混合难题	reinforcement learning
5	Learning Controllable and Diverse Player Behaviors in Multi-Agent Environments	提出一种可控且多样化的多智能体行为学习框架，用于游戏AI。	reinforcement learning PPO

#	题目	一句话要点	标签
6	Limits and Gains of Test-Time Scaling in Vision-Language Reasoning	研究测试时缩放（TTS）在视觉-语言推理中的局限与收益	large language model multimodal
7	TAO-Net: Two-stage Adaptive OOD Classification Network for Fine-grained Encrypted Traffic Classification	提出TAO-Net，用于细粒度加密流量分类中的未知流量识别与分类。	large language model
8	GPG: Generalized Policy Gradient Theorem for Transformer-based Policies	提出Transformer策略的广义策略梯度定理，为LLM高效优化提供新视角	large language model
9	CIEGAD: Cluster-Conditioned Interpolative and Extrapolative Framework for Geometry-Aware and Domain-Aligned Data Augmentation	CIEGAD：一种聚类条件下的几何感知和域对齐数据增强框架，用于解决数据稀疏和类别不平衡问题。	large language model