Chapter 7. Reinforcement Learning

Let's talk about the nature of learning. We aren't born into this world knowing anything. By interacting with the world, we learn about the effects of our actions. Once we have an understanding of how the world works, we can use that knowledge to make decisions that can lead us to specific goals.

In this chapter, we will formulate this approach to learning computationally using a method called reinforcement learning. It's very different to the other types of deep learning algorithms covered in this book and is a vast field on its own.

Applications of reinforcement learning range from playing games in a digital environment to governing the actions of robots in a real-life environment. It also happens to be the technique you use to train dogs and other animals. These days, reinforcement learning is being used to drive self-driving cars and is a hugely popular field.

One of the major recent breakthroughs happened when a computer...