Basics of Deep Learning
We have already implemented deep learning algorithms in Chapter 03, Deep Learning in Practice using TensorFlow 2. Before we begin with deep Q learning, which is the focus of this chapter, it is essential that we quickly revise the basics of deep learning.
Let us first understand what a perceptron is before we look into neural networks. The following figure represents a general perceptron:
Figure 9.1: Perceptron
A perceptron is a binary linear classifier, where the inputs are first multiplied by the weights, and then we take a weighted sum of all these multiplied values. Then, we pass this weighted sum through an activation function or step function. The activation function is used to convert the input values into certain values, such as (0,1), as output for binary classification. This whole process can be visualized in the preceding figure.
Deep feedforward networks, which we also refer to as Multilayer Perceptrons (MLPs...
