Let's examine the structure and operation of a simple neural network. This will only be a brief overview and will serve as a transition to the deep Q-network that we will be constructing later in this chapter. If you are already familiar with how basic networks and activation functions work, feel free to skip to the Building your first deep Q-network section.

We will not be doing an exhaustive primer on neural networks, but it's worth researching them further and becoming familiar with the basic concepts of propagating values through layers to make predictions and backpropagating errors to update the network based on the results. We'll provide an overview of this process here.