发帖
雷达卡
Reinforcement Learning with Prototypical Representations
Denis Yarats 1 2 Rob Fergus 1 Alessandro Lazaric 2 Lerrel Pinto 1
Abstract from rewards alone is sample inefficient and leads to poor
Learning effective representations in image-based performance. Prior work (Srinivas et al., 2020; Yarats et al.,
environments is crucial for sample efficient Rein- 2019) addresses this problem by leveraging self-supervised
forcement Learning (RL). ...
京公网安备 11010802022788号
