Making Decisions in Complex Environments with Reinforcement Learning

In the previous chapter, we focused on RNNs for sequential learning. The last chapter of the book will be about reinforcement learning, which is the third type of machine learning task mentioned at the beginning of the book. You will see how learning from experience and learning by interacting with the environment differs from previously covered supervised and unsupervised learning.

We will cover the following topics in this chapter:

• Setting up a workspace for reinforcement learning
• Basics of reinforcement learning
• Simulation of OpenAI Gym environments
• Value iteration and policy iteration algorithms
• Monte Carlo methods for policy evaluation and control
• The Q-learning algorithm