Our work “Necessary and Sufficient Conditions for Inverse Reinforcement Learning of Bayesian Stopping Time Problems” got accepted for publication in the Journal of Machine Learning Research (JMLR). This paper extends inverse reinforcement learning to Bayesian stopping time problems where the inverse learner has no knowledge of the decision maker’s model dynamics or private observations.
In spite of a 2.5-year long arduous review period, the 55 page manuscript is set to appear in JMLR, 2023.