← the wiki

Machine Learning

Programs that improve from examples instead of being given explicit rules for every case.

Classic programming: a human writes the rules, the machine applies them. Machine learning flips it: you show the machine thousands of examples — emails labeled spam/not-spam, photos labeled cat/dog — and it works out the rules itself.

The result is a model: a program grown from data rather than written by hand. It shines exactly where humans can't write the rules down — recognizing a face, translating a sentence, predicting what you'll type next.

The best way to make it stick: watch it run. Start an interactive lesson →