Chapter 3 · Neural networks
Layers and the multi-layer network
Stacking layers lets networks build complex features from simple ones.
Preparing layers and the multi-layer network
Chapter 3 · Neural networks
Stacking layers lets networks build complex features from simple ones.
Raw numbers go in
Each layer takes the numbers from the layer before, mixes them and passes new numbers on. Early layers spot simple things; later layers combine them.
A layer is a group of neurons that all look at the same inputs at the same time. Each one makes its own small decision.
Pick a puzzle, choose how many hidden layers and neurons, then press Train. The map shows what the network believes everywhere on the page.
Opposite corners match (the XOR puzzle).
Press Train to start.
What does the second layer receive as its input?
Why add a hidden layer?
What is a hidden layer?