GAN (Generative Adversarial Network)

Introduced by Goodfellow et al. (2014), a GAN pits a generator that produces synthetic samples against a discriminator that judges real versus fake. As each improves, the generator learns to produce increasingly realistic output. GANs drove a generation of photorealistic image synthesis.

For image generation they have largely been succeeded by diffusion models, but the adversarial-training idea remains foundational.