Aravind Srinivas
TalkRL: The Reinforcement Learning Podcast
00:00
Contrastive Learning in Reinforcement Learning
This chapter explores the use of contrastive learning in reinforcement learning models to improve data efficiency and achieve better results. They discuss the concept of instance discrimination and how data augmentations and contrastive losses are incorporated in the latent space. The chapter also compares this approach to other methods and discusses its potential applications.
Transcript
Play full episode
