Reinforcement learning is about developing goal-driven agents to automate problem-solving by optimizing their actions within an environment. This involves predicting and classifying the available data and training agents to execute tasks successfully. Generally, an agent is an entity that has the capacity to interact with an environment, and the learning is done by applying feedback in terms of cumulative rewards from the environment to inform future actions.

Three different types of reinforcement learning can be distinguished:

• Value-based—a value function provides an estimate of how good the current state of the environment is.
• Policy-based—where a function determines an action based on a state.
• Model-based—a model of the environment including state transitions, rewards, and action planning.

In this chapter, we'll start...