View of KL-Regularized Reinforcement Learning is Designed to Mode Collapse

##article.return## KL-Regularized Reinforcement Learning is Designed to Mode Collapse Download Download PDF