arXiv Robotics — research abstracts· Zhilin He, Gauri Joshi·· 2 days agoEditorial score42
FedGuide: A Framework for Heterogeneous Federated Reinforcement Learning
FedGuide: Diffusion Prior Alignment and Value Baseline Guidance for Heterogeneous Federated Reinforcement Learning
Summary
FedGuide improves FRL by aligning diffusion priors and using DICE for value guidance. It outperforms existing methods in returns and robustness across diverse environments.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org