13. Deep Reinforcement Learning

In Chapter 4, we introduced the paradigm of reinforcement learning (as distinct from supervised and unsupervised learning), in which an agent (e.g., an algorithm) takes sequential actions within an environment. The environments—whether they be simulated or real world—can be extremely complex and rapidly changing, requiring sophisticated agents that can adapt appropriately in order to succeed at fulfilling their objective. Today, many of the most prolific reinforcement learning agents involve an artificial neural network, making them deep reinforcement learning algorithms.

In this chapter, we will

  • Cover the essential theory of reinforcement learning in general and, in particular, a deep reinforcement learning ...

Get Deep Learning Illustrated: A Visual, Interactive Guide to Artificial Intelligence now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.