Persistent advantage learning dueling double DQN for the Arcade Learning Environment - View it on GitHub