Short Definition
Unsupervised learning finds structure in data without using explicit target labels.
Intuition
Instead of learning from an answer key, the model looks for patterns such as clusters, dimensions, topics, or compressed representations.
Technical Definition
Unsupervised learning optimizes objectives that expose latent structure in input data, such as clustering, density estimation, dimensionality reduction, or representation learning.
Example
Customer segments can be found by clustering purchasing behavior without predefining the segment labels.
Common Misunderstandings
Unsupervised does not mean unguided; the objective still shapes what the model learns.
Patterns found without labels still need validation before they are trusted.