Backpropagation

Backpropagation is how a network finds the slope of the loss for every weight at once. It runs the network forward to get a prediction and a loss, then works backwards layer by layer, using the chain rule to pass along how much each weight contributed to the error. Gradient descent then uses those slopes to update all the weights.

Why it matters: It's what makes training networks with billions of weights possible.

Builds on: The chain rule, Gradient descent, Layers and networks

Learn it, then prove it

Sign in (free) to take this step's mastery check and save your journey.