BriefGPT.xyz
Sep, 2023
关于最优传输在课程强化学习中的益处
On the Benefit of Optimal Transport for Curriculum Reinforcement Learning
HTML
PDF
Pascal Klink, Carlo D'Eramo, Jan Peters, Joni Pajarinen
TL;DR
通过将课程设置为任务分布之间的插值,将生成课程作为约束优化传输问题来提高课程强化学习(CRL)方法的性能,从而在具有不同特点的各种任务中取得高性能。
Abstract
curriculum reinforcement learning
(CRL) allows solving complex tasks by generating a tailored sequence of learning tasks, starting from easy ones and subsequently increasing their difficulty. Although the potential of
c
→