Math for Deep Learning

2PROBABILITY

Probability affects every aspect of our lives, but in reality, we’re all pretty bad at it, as some of the examples in this chapter demonstrate. We need to study probability to get it right. And we need to get it right because deep learning deals extensively with ideas from probability theory. Probability appears everywhere, from the outputs of neural networks to how often different classes appear in the wild to the distributions used to initialize deep networks.

This chapter aims to expose you to the sorts of probability-related ideas and terms you’ll frequently encounter in deep learning. We’ll start with basic ideas about probability ...