Pong played by an AI using deep reinforcement learning

Keras/Tensorflow were utilized to model the neural network for the Deep Q-Learning algorithm.

After the episodes are finished, results can be found on "pong.png".

Models will be saved each time the last average score is beaten.