ContentsThe library

Backpropagation in Full

A Product of Many Small Numbers

Last timeThe Memory the Backward Pass Costs

The derivative reaching an early layer is a product of one factor per layer in between. Products of many numbers go to zero or to infinity, and almost every architectural trick is a response to that.

Everything in this course has concerned one operation at a time. This last lesson

is about what happens when there are hundreds of them in a row, and it is a

statement about products rather than about networks.

One factor per layer

Apply the chain rule down the depth of a network and the derivative reaching an

early layer is a product with one term for every layer above it. That is all.

FIG 1The derivative at an early layer
the local factor contributed by layer k, roughly how much it scales a small change
the number of layers between here and the loss
the derivative that finally arrives at the early layer
Nothing in this expression is specific to neural networks. It is the chain rule applied many times, and the whole of what follows is the behaviour of a product of many numbers.

Why a hair matters

Products of numbers near one are unforgiving.

The lesson stops here

5 more paragraphs to go

You have read the opening. The rest of the argument, the problems that check whether it landed, and the lines worth keeping at the end all come with a plan.

The first lesson of every course in the library reads the whole way through, free, so you can see exactly what the rest of them are.

See the planThe contents

This is the reading half

Starting the course gives you your own copy of it. Every idea on every page has problems standing under it, marked with a reason rather than a tick, and any sentence you do not believe can be opened and argued with. None of that can happen on a page nobody owns.

The contents

The rest of this course

  1. 01There Are No Layers, Only Operations
  2. 02Multiply Along, Add Acrossopening only
  3. 03Two Ways to Accumulate, and Only One Is Affordableopening only
  4. 04Four Rules That Cover Most of a Networkopening only
  5. 05The Matrix Everybody Writes Down and Nobody Buildsopening only
  6. 06Why the Last Two Steps Are One Stepopening only
  7. 07What the Forward Pass Has to Leave Behindopening only
  8. 08A Product of Many Small Numbersyou are here

Read alongside