Overview / Usage

I have build the game of Ping Pong (http://www.ponggame.org) from scratch. Then I build a Deep Q Network that gets better and better over time through trial and error. The DQN is a convolutional neural network that reads in pixel data from the game and the game score. Using just those 2 parameters, it learns what moves it needs to make to become better.
This project is open to any contribution that can enhance the training process or you can add some cool features to the pong game.