Generative adversarial networks

This class of machine learning systems, known as Generative adversarial networks (GANs), comprises two networks. One is the generator, responsible for creating new data samples, and the other is the discriminator, which differentiates between the generated and real data. The two networks undergo simultaneous training, with the generator striving to produce data that the discriminator cannot differentiate from real data, and the discriminator aiming to improve at distinguishing real data from the generated data. This adversarial process results in the generator network producing high-quality data.