发帖
雷达卡
Unsupervised Curricula
for Visual Meta-Reinforcement Learning
Allan Jabriα Kyle Hsuβ,0 Benjamin Eysenbachγ
Abhishek Guptaα Sergey Levineα Chelsea Finnδ
α
UC Berkeley β University of Toronto γ Carnegie Mellon University δ Stanford University
Abstract
In principle, meta-reinforcement learning algorithms leverage experience across
many tasks to learn fast and effective reinforcement learning (RL) strategies. How ...
京公网安备 11010802022788号
