Unsupervised learning can be used in conjunction with Bayesian inference to produce conditional probabilities (i.e., supervised learning) for any of the random variables given the others.

Unsupervised learning is also useful for data compression: fundamentally, all data compression algorithms either explicitly or implicitly rely on a probability distribution over a set of inputs.

Another form of unsupervised learning is clustering, which is sometimes not probabilistic. Also see formal concept analysis.