This is a backgammon AI that can be trained on self-play based on a neural network. The neural network architecture can be modified in the "player.py" files
Note: the doubling cube has not been implemented yet
Explanation of RL: https://keon.io/deep-q-learning
Backgammon paper: https://bkgm.com/articles/tesauro/ProgrammingBackgammon.pdf