Aravind Srinivas
TalkRL: The Reinforcement Learning Podcast
00:00
Contrastive Predictive Coding (CPC) and its Applications
The chapter discusses the concept of Contrastive Predictive Coding (CPC) and its applications as a latent space generative model trained with contrastive losses. The speakers highlight the importance of self-supervised learning and unsupervised learning in the context of CPC. They also introduce the concept of Contrastive Unsupervised Representations for Reinforcement Learning (CURL).
Transcript
Play full episode
