Supervised vs. Unsupervised Learning

Learning from labeled examples versus finding structure in unlabeled data — and where semi-supervised and reinforcement learning fit.

What it is

Supervised learning trains a model on (input, label) pairs to predict labels for new inputs — classification and regression are the two main flavors. Unsupervised learning works on unlabeled data, finding structure directly — clustering, dimensionality reduction, and anomaly detection.

Key points

Supervised: needs labeled data, optimizes directly against a known target, easiest to evaluate objectively.
Unsupervised: no labels required, evaluation is harder (often indirect, e.g. cluster cohesion), useful when labels are expensive or the goal is exploratory.
Semi-supervised and self-supervised learning sit between the two — using a small labeled set or a pretext task to bootstrap from mostly unlabeled data, which is how most modern foundation models are pretrained.
Reinforcement learning is a distinct third paradigm: an agent learns from reward signals through interaction, not from a fixed dataset.