All terms
Architectures
Generative Adversarial Network
Two networks trained against each other: one creates fake samples, the other tries to detect them.
Definition
A generative adversarial network pairs a generator that creates synthetic samples from random noise with a discriminator that tries to tell real data from generated data. The two improve by competing in a minimax game, pushing the generator toward increasingly realistic output. GANs drove early breakthroughs in photorealistic image synthesis and remain useful for fast, high-resolution generation, though diffusion models have largely surpassed them on image quality.