https://medium.com/@monocosmo77/how-inverse-reinforcement-learning-works-part4-artificial-intelligence-615d8a8a901b