Generative Adversarial Network

A machine learning approach in which two competing neural networks are given a training set and through competition create a new data set with the same statistical attributes as the training set.