École Nationale Supérieure de
Techniques Avancées

Abstract

This paper discusses a literature review on generative adversarial networks (GANs). Thispaper will first develop the architecture of GANs, by presenting the principle of this generativemodel as well as the generator network and the discriminator network, and then the theory, byfocusing on the notions of machine learning, unsupervised learning, game theory, random vectorgeneration method, implementation of cost function and by exploiting the theoretical results.Finally, the limitations of this model will be presented and the uses and improvements that havebeen made to GANs will be discussed.