Clustering is a core task within unsupervised learning where the goal is to partition an unlabeled dataset into groups of similar objects or data points. Clustering aims to find the natural groupings or patterns within data without relying on predefined labels.

In practical terms, clustering translates into grouping data points based on their similarity, such as segmenting articles on a website based on their topic (topic modeling). In some algorithms, the user may define the number of clusters or the characteristics used for grouping, but the validation of the output relies on external evaluation and interpretability.