Distilling a Hierarchical Policy for Planning and Control via Representation and Reinforcement Learning https://arxiv.org/abs/2011.08345