Concept guide · Foundations
What is unsupervised learning?
A slower explanation for when the short interview answer is not enough. Start with the intuition, then connect it to the technical idea and the way an interviewer expects you to reason about it.
ML area
Foundations → Unsupervised learning
Interview format
Foundations
Question
ML-004
Start with the idea
In simple terms
Unsupervised learning looks for useful structure when the dataset does not provide a correct answer for every example.
Make it concrete
A retailer might group customers by similar behaviour even when nobody has labelled the groups in advance.
Build the answer
How to reason about it
- 1Confirm that there is no supplied target used as the answer.
- 2Name the kind of structure being discovered.
- 3Explain how you would check whether that structure is useful or stable.
Go one level deeper
The technical explanation
The structure may be clusters, a lower-dimensional representation, or unusual observations. Because there is no supplied target, evaluation is less direct: you need to ask whether the discovered structure is stable, interpretable, and useful for a downstream decision. A cluster is not automatically a real business segment, and an anomaly score is not automatically proof that something is fraudulent.
Common mistake
Using clustering as if it were the definition of all unsupervised learning.
Say this in the interview
Unsupervised learning works without a supplied target. The model looks for structure in the inputs, such as groups, lower-dimensional representations, or unusual points. Clustering is one example, but dimensionality reduction and anomaly detection are also unsupervised. The output is usually a tool for discovery or downstream decisions, not automatically a ground-truth label.
Keep preparing