Hands-On Q-Learning with Python

Perceptron functions

The simplest kind of neural network to understand is a binary classifier called the perceptron. The following diagram is an overview of how the perceptron works:

We see that the inputs are coming in and being multiplied by the weights, and then an overall weighted sum is taken:

This is the process of taking the dot product of two matrices or vectors. The values at the weighted sum node are then sent to an activation function called a step function

An activation function effectively maps inputs onto outputs and determines ...