In this section, we will look at how to implement a simple reinforcement learning method. We will use OpenAI's open source toolkits, gym and universe, for this purpose. gym is a software framework for developing and comparing different reinforcement learning algorithms. It supports environments for different games by Atari, board games, as well as classic control tasks. universe on the other hand provides a wrapper on top of gym with modules for client and server, where you can visualize the progress of a reinforcement learning system. To install these toolkits on a macbook, we do the following:
pip install gym
brew install golang libjpeg-turbo
pip install universe
