Foundations · 1 min read · Updated 2026-10-05
Types of machine learning at a glance
Supervised: learn with answers
The computer gets examples that already have the right answer. A spam filter learns from emails marked 'spam' or 'not spam'. Another example is guessing house prices from past sales.
Unsupervised: find your own groups
The computer gets data with no answers and looks for patterns. A shop can use it to find groups of customers who buy similar things.
Reinforcement: learn from rewards
The computer tries moves and gets points. It repeats what earns points. This is how programs learn to play games well.
Key takeaways
- Supervised has answers, unsupervised has none, reinforcement has rewards.
- Each type suits a different kind of problem.
- Chatbots use a mix of all three.
Quick questions
Which type is most common?
Supervised learning is used a lot, because many problems come with examples and answers.
Can one AI use more than one type?
Yes. Chatbots use all three at different steps of their training.
Which is the hardest?
Reinforcement learning is often harder to set up, because you must design good rewards.
Keep learning
Supervised learningSupervised learning is learning with an answer key. The computer sees many examples where the right answer is already…1 min readRead →Unsupervised learningUnsupervised learning is learning with no answer key. The computer looks at lots of data and finds groups and patterns on…1 min readRead →Reinforcement learningReinforcement learning is learning by trying things and getting rewards. The computer repeats what earns points and avoids…1 min readRead →