Imitation Learning by Reinforcement Learning

Kamil Ciosek

Imitation learning algorithms learn a policy from demonstrations of expert behavior. We show that, for deterministic experts, imitation learning can be done by reduction to reinfor…