In this chapter, we learned that deep learning is differentiated from other machine-learning algorithms because of the use of multiple hidden layers. This network of hidden layers, which are composed of artificial neurons, was designed to mimic the way our brain processes input signals to interpret our environment. The units within the hidden layers take in all the independent variables and apply some weights to these variables. In this way, each neuron classifies the combination of input values in different ways.
From understanding the architecture of this type of machine learning from a high level, we then took a deeper dive into the actual process of converting the input to predictions using this approach. We discussed the various activation functions that act as the gate for every neuron, determining whether a signal should be passed to the next layer. We then built...
