In this chapter, I'm going to briefly introduce deep learning with some examples based on Keras. This topic is quite complex and needs dedicated books; however, my goal is to allow the reader to understand some basic concepts that can be helpful before starting a complete course. In the first section, I'm presenting the structure of artificial neural networks and how they can be transformed in a complex computational graph with several different layers. In the second one, I'm going to introduce the basic concepts of Keras, and we'll see an example based on a very famous test dataset.
In particular, we are going to discuss the following:
- Structure of an artificial neuron
- Multi-layer Perceptrons (MLP)
- The back propagation algorithm
- How to build and train an MLP with Keras
- How to interface Keras with scikit-learn
