GAN

Generative Adversarial Network

Two-network setup -- generator vs. discriminator -- for synthesizing data.

When you'd see it: Image and media synthesis discussions. GANs powered an early wave of realistic generated faces and deepfakes.

Why it matters: A GAN pits two networks against each other, one generating fakes and one trying to catch them, until the fakes are convincing. It was a breakthrough for generated images, though diffusion models now lead much of that work.