# Active Learning

> Source: https://sukruyusufkaya.com/en/glossary/aktif-ogrenme
> Updated: 2026-05-13T21:12:10.636Z
> Type: glossary
> Category: yapay-zeka-temelleri
**TLDR:** A data-efficiency approach in which the model selects the most informative examples and requests labels from a human or expert source.

<p>Active learning offers a smart strategy, especially in domains where labeling is expensive. Instead of treating all data as equally useful, the model selects the examples that are most valuable for improving itself and requests labels only for those. This makes it possible to reach higher learning quality with less human effort. In fields such as medicine, law, defense, and expert annotation workflows, active learning can create major efficiency gains. Its value comes not from using more data, but from identifying which data matters most.</p>