https://medium.com/@monocosmo77/how-inverse-reinforcement-learning-works-part3-artificial-intelligence-8a9520834f78