The Japanese Society for Artificial Intelligence

Recently, reinforcement learning (RL) has been showing increasingly high performance in a variety of complex tasks of decision making and control, but RL requires quite careful engineering of reward functions to solve real tasks. Inverse reinforcement learning (IRL) is a framework to construct reward functions by learning from demonstration, but there is no way to guarantee the performance of the learned reward functions in maximum entropy IRL, the mainstream of IRL. Therefore it is unclear how reliable the results can be. To provide a theoretical guarantee on the performance of maximum entropy IRL, we evaluate and discuss its performance theoretically.