Reasoning about Counterfactuals to Improve Human Inverse Reinforcement Learning | IEEE Conference Publication | IEEE Xplore