GAN stands for generative adversarial network. It is a type of a CNN neural network architecture in which two types of CNN neural networks compete against each other. In a GAN network architecture there are typically one or more generator neural networks and one ore more discriminator neural networks. The generator network performs continuous iterations, in an attempt to build/generate content (usually an image) from pure noise and by using datasets as training data. At the same time, a discriminator neural network takes the generator network as input and is trained on a series of images which are real images, against which the generator competes. The discriminator is essentially a classifier, which outputs a boolean value of true or false, which corresponds to whether or not the generator network created image is a forgery of a real image or not.
