Intermediate · Fundamentals

Semi-supervised Learning

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. A hybrid approach that uses a small amount of labeled data alongside a large amount of unlabeled data.

Technical Definition

A hybrid approach that uses a small amount of labeled data alongside a large amount of unlabeled data.

How it works

Labeling data is expensive; unlabeled data is abundant. Semi-supervised methods leverage both — for example, using a model trained on the labeled set to pseudo-label the unlabeled set, then retraining on the combined data. Techniques like consistency regularization and self-training have closed much of the gap with fully supervised models.

Related Concepts

• Supervised Learning — Learning from input–output pairs where each training example carries a correct label.
• Unsupervised Learning — Learning patterns from data that has no labels — only the inputs.
• Label — The correct answer attached to a training example in supervised learning.