Fundamentals
Unsupervised learning
Unsupervised learning finds structure in data without supplied target labels.
How it works
The objective may group examples, compress them, or estimate their distribution. The result is useful structure rather than a direct answer key.
A concrete example
Illustrative: cluster 12,000 customer vectors into k = 8 groups.
Do not confuse it with Self-supervised learning
self-supervision constructs targets from the data itself.