Generative adversarial network (GAN)

A generative adversarial network (GAN) is a class of machine learning systems where two neural networks contest with each other in a game (in the sense of game theory, often but not always in the form of a zero-sum game). Given a training set, GAN learns to generate new data with the same statistics as the training set. For example, a GAN trained on photographs can generate new photographs that look at least superficially authentic to human observers, having many realistic characteristics. Though originally proposed as a form of generative model for unsupervised learning, GANs have also proven useful for semi-supervised learning, fully supervised learning, and reinforcement learning.

https://medium.com/@nutorbitx/gans-%E0%B8%AD%E0%B8%B0%E0%B9%84%E0%B8%A3%E0%B8%84%E0%B8%B7%E0%B8%AD-generative-adversarial-networks-7973ae70db70

The well-known application of GAN is deepfake.