Skip to content

Foundations

Derivatives and gradients

Training a model means walking downhill on its loss, and the slope says which way is down. This topic reads the derivative as the slope at a point, gathers the slopes over two parameters into the gradient, and passes rates along a chain of functions with the chain rule.

Lessons on this topic

Beginner · 25 minYou are hereRead

The derivative as a slope

The derivative is the slope of a curve at a point, the limit of slopes between two ever closer points; at the bottom of a smooth valley it is zero.

Intermediate · 30 minYou are hereRead

Partial derivatives and the gradient

On a surface over two parameters, take the slope along each axis with the other held still; the two slopes form the gradient, which points straight uphill.

Phasorium

Account

Sign inKeep your place on any device

Subjects

Site

Look