Chapter 10 — Optimisation and Backpropagation

(Roadmap — content to be written)

This chapter covers:

• Loss functions and gradient descent

• The chain rule as a graph traversal

• Computational graphs and automatic differentiation

• Backpropagation through a convolution layer

• Learning rate, momentum, Adam

Depends on: Chapter 9 Builds toward: Chapter 11 (CNNs)