Recent Advance On Generative Adversarial Networks

Abstract

Generative adversarial networks (GANs) has received wide attention in the machine learning field because it can generate real-like data by estimating real data probability distribution. GANs has been successfully applied to many fields such as computer vision, pattern recognition, natural language processing and so on. By now many kinds of extended models of GANs have been proposed and investigated by different researchers from different viewpoints. Although there are a few review papers on the extended models of GANs in the literature, some remarkable extensions of GANs published in the recent years are not included in these surveys. This paper attempts to provide the potential readers with a recent advance on GANs by surveying its twelve representative variants. Furthermore, we also present the lineage of the extended models of GANs. This paper can provide researchers engaged in related works with very valuable help.