Adversarial Learning

A technique where an AI agent learns by competing against another model (often called a "discriminator").

In adversarial training, particularly with Generative Adversarial Networks (GANs), one model generates data while the other evaluates it, forcing the generator to improve.

This is commonly used in generating high-quality images, videos, and in some cybersecurity applications.