rl-bubble-breaker Reinforcement Learning of Bubble Breaker in TF/Keras and Theano/Lasagne There's also a presentation about the background to this that I gave in Korea