RL is the machine learning technology currently dominating the interest of many researchers. It is typically appealing to us, because it fits well with games and simulations. In this chapter, we covered some of the foundations of RL by starting with the fundamental introductory problems of the multi-armed and contextual bandits. Then, we quickly looked at installing the OpenAI Gym RL toolkit. We then looked at Q-learning and how to implement that in code and train it on an OpenAI Gym environment. Finally, we looked at how we could conduct various other experiments with Gym by loading a couple of other environments, including the Atari games simulator.

In the next chapter, we look at the quickly evolving a cutting-edge RL platform that Unity is currently developing.