On the Importance of Feature Decorrelation for Unsupervised Representation Learning in Reinforcement Learning

Open in new window