Summary
In this chapter, we learned about a whole new field of unsupervised learning: reinforcement learning. It is a whole different field and we have just touched on this topic in this chapter. We learned how to phrase a problem for reinforcement learning, and then we trained a model that sees a few measurements provided by the environment and can learn how to balance a cartpole. You can apply the same knowledge to teach robots to walk, to drive cars, and also to play games. This is one of the more physical applications of deep learning.
In the next and closing chapter, we'll be looking at productionizing our PyTorch models so that you can run them on any framework or language, and scale your deep learning applications.
