Clustering

Grouping items by similarity with no labels to learn from. Useful for exploring a dataset, and treacherous as a conclusion, because any method will return groups whether or not the data has any.

A definition is not an understanding. This is taught properly in Finding structure without labels, part of Unsupervised learning.