Intermediate · Research

Active Learning

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

TL;DR. An ML approach where the model selects which examples it most wants labeled next.

Technical Definition

An ML approach where the model selects which examples it most wants labeled next.

How it works

Instead of labeling random data, active learning picks the most informative examples — typically those near the model's decision boundary or with high predictive uncertainty. This can dramatically reduce labeling cost in domains where labels are expensive (medical, legal).

Related Concepts

• Semi-supervised Learning — A hybrid approach that uses a small amount of labeled data alongside a large amount of unlabeled data.
• Annotation — The act of attaching labels, tags, or structured information to raw data.
• Human-in-the-loop — System designs where humans review, correct, or guide AI outputs as part of the workflow.