# Supervised, Semi-supervised, Self-supervised: How Labeling Needs Change Across Paradigms

> Source: https://sukruyusufkaya.com/en/learn/veri-etiketleme-kalite/supervised-semi-self-supervised-veri-etiketleme
> Updated: 2026-05-13T13:03:22.514Z
> Category: Veri Etiketleme & Kalite Yönetimi
> Module: Module 1: The Anatomy of Data Labeling
**TLDR:** Modern AI's five major learning paradigms — supervised, semi-supervised, self-supervised, weakly supervised, and few-shot/in-context — each with its labeling needs, cost profile, and use cases.