Skip to content

Foundations

Learning from data

Machine learning chooses a rule from examples instead of writing it by hand. This topic turns examples into points in feature space, measures how wrong a model is with a loss, and makes a first prediction from a point's nearest neighbours.

Lessons on this topic

Beginner · 25 minYou are hereRead

What machine learning is

Instead of writing a rule by hand, pick one from a family of rules by how well it fits examples, and judge it on new inputs.

Beginner · 25 minYou are hereRead

Models and loss

A loss turns how wrong a model is into one number; fitting means making it small, and the choice of loss changes which model wins.

Beginner · 25 minYou are hereRead

Nearest neighbours

Label a new point by asking the training examples nearest to it, and see how the number asked decides how smooth the answer is.

Related topics

Where to go next.

Phasorium

Account

Sign inKeep your place on any device

Subjects

Site

Look