Math for Deep Learning

9DATA FLOW IN NEURAL NETWORKS

In this chapter, I’ll present how data flows through a trained neural network. In other words, we’ll look at how to go from an input vector or tensor to the output, and the form the data takes along the way. If you’re already familiar with how neural networks function, great, but if not, walking through how data flows from layer to layer will help you build an understanding of the processes involved.

First, we’ll look at how we represent data in two different kinds of networks. Then, we’ll work through a traditional feedforward network to give ourselves a solid foundation. We’ll see just how compact inference with ...