Exploiting Hierarchy for Learning and Transfer in KL-regularized RL

Open in new window