A Survey of Inverse Reinforcement Learning: Challenges, Methods and Progress - 2021

Title : A Survey of Inverse Reinforcement Learning: Challenges, Methods and Progress Author(s): Arora, Saurabh and Doshi, Prashant Link(s) :

Survey on Inverse Reinforcement Learning (IRL).

The problem is for the agent (learner) to infer the reward function given trajectories from an agent (expert) who is assumed to have followed an optimal policy to generate these trajectories.