BriefGPT.xyz
Jan, 2024
将离线强化学习重新构建为回归问题
Reframing Offline Reinforcement Learning as a Regression Problem
HTML
PDF
Prajwal Koirala, Cody Fleming
TL;DR
该研究将离线强化学习重新定义为一个可以用决策树解决的回归问题，通过梯度提升树可以实现快速训练和推理，同时对通用性进行了讨论。
Abstract
The study proposes the reformulation of
offline reinforcement learning
as a
regression problem
that can be solved with
decision trees
. Aim
→