https://medium.com/@monocosmo77/how-deep-reinforcement-learning-drl-progressed-in-2022-part2-668e4fa0dcd6