Foundations · 1 min read · Updated 2026-10-05
Unsupervised learning
How it works
We give the computer data but no right answers. It looks for things that are similar and puts them together. A shop might use it to find groups of customers who buy similar things.
What it is good for
Finding groups, spotting odd things that do not fit (like a strange card payment), and shrinking big data into something easier to understand.
Why it matters for chatbots
There is far more text without answers than with answers. A trick called self-supervised learning lets the computer make its own questions from plain text. That is how chatbots learn language.
Key takeaways
- No answers are given. The computer finds patterns itself.
- Good for grouping things and spotting odd ones.
- Learning from plain text made chatbots possible.
Quick questions
What is clustering?
Putting similar things into the same group.
How is it different from supervised learning?
Supervised learning has an answer key. Unsupervised learning has none.
What is self-supervised learning?
The computer hides a word in a sentence and tries to guess it. The sentence itself is the answer key.