https://medium.com/@monocosmo77/how-offline-reinforcement-learning-works-part1-artificial-intelligence-d72864d3bd8c