Active Learning

Intermediate

Selecting the most informative samples to label (e.g., uncertainty sampling) to reduce labeling cost.

Full Definition

Selecting the most informative samples to label (e.g., uncertainty sampling) to reduce labeling cost.

Keywords

query strategy

Domains

Foundations & Theory

Related Terms

Curriculum Learning related to

Ordering training samples from easier to harder to improve convergence or generalization.

 Data Augmentation related to

Expanding training data via transformations (flips, noise, paraphrases) to improve robustness.

 Synthetic Data related to

Artificially created data used to train/test models; helpful for privacy and coverage, risky if unrealistic.

 Federated Learning related to

Training across many devices/silos without centralizing raw data; aggregates updates, not data.

 Confounding related to

A hidden variable influences both cause and effect, biasing naive estimates of causal impact.

 A/B Testing related to

Controlled experiment comparing variants by random assignment to estimate causal effects of changes.

 Data Labeling related to

Human or automated process of assigning targets; quality, consistency, and guidelines matter heavily.

 Inter-Annotator Agreement related to

Measure of consistency across labelers; low agreement indicates ambiguous tasks or poor guidelines.

Concept Map

See how Active Learning connects to other concepts.

Open Knowledge Graph