VOCE: Variational Optimization with Conservative Estimation for Offline Safe Reinforcement Learning

Open in new window