Foundations · 1 min read · Updated 2026-10-05
Supervised learning
How it works
Say we want to find spam emails. We show the computer thousands of emails, each marked 'spam' or 'not spam'. It guesses, checks the answer, and fixes its mistakes. After many rounds it gets good at it.
Two common jobs
Sorting things into groups, such as spam or not spam, or cat or dog. And guessing a number, such as the price of a house.
The catch
Someone has to mark all those examples first. That takes time and money, and wrong marks teach the computer wrong things.
Key takeaways
- The computer learns from examples with answers.
- It can sort things into groups or guess numbers.
- Getting good answer labels is the hard part.
Quick questions
What is a label?
It is the right answer attached to an example, like the word 'spam' on a junk email.
What is overfitting?
It is when the computer memorises the practice examples but fails on new ones, like a student who learns the answers by heart but cannot solve a new question.
Are chatbots supervised?
Partly. They first learn from plain text, then from examples of good answers.