Supervised learning uses labeled data where the model learns from input-output pairs (e.g., classification, regression). Unsupervised learning uses unlabeled data to find patterns and structures (e.g., clustering, dimensionality reduction).