AlgoPlus//machine learning / backprop
Read the theory

Backpropagation

The chain rule sends the loss gradient backward through every layer.

Direction
Loss
Legend
Forward activation
Backward gradient
AI Tutor Workspace
In a nutshell
Backpropagation is how a network learns from a mistake. A forward pass makes a prediction; the loss measures how wrong it is. Backprop then works backward from the output, using the chain rule to hand every weight a gradient — the direction and size of the nudge that would shrink the loss. An optimizer applies those nudges, and that one round trip is a single training step. It costs about the same as a forward pass.
Ready
Press play to begin the cinematic walkthrough.
Blame, distributed fairly. The output's error is split back across the neurons that caused it, layer by layer, each getting exactly the share the chain rule assigns.
Key terms
Go deeper in the lesson
Read the full theory, intuition & complexity for Backpropagation.