A model architecture where two neural networks (a generator and a discriminator) compete, producing increasingly realistic synthetic data.
Friendly Description: A GAN is a clever AI setup with two networks playing a friendly game against each other. One network (the "generator") tries to create fake images that look real, and the other (the "discriminator") tries to spot the fakes. Each one keeps getting better by trying to outsmart the other, and the result is shockingly realistic generated content.
Example: GANs are how some early apps could create realistic photos of people who don't exist, or turn a daytime photo into a convincing nighttime scene. The two networks effectively coach each other through millions of rounds until the fakes are nearly indistinguishable from real photos.