Some Theoretical Properties of GANs

Abstract : Generative Adversarial Networks (GANs) are a class of generative algorithms that have been shown to produce state-of-the art samples, especially in the domain of image creation. The fundamental principle of GANs is to approximate the unknown distribution of a given data set by optimizing an objective function through an adversarial game between a family of generators and a family of discriminators. In this paper, we offer a better theoretical understanding of GANs by analyzing some of their mathematical and statistical properties. We study the deep connection between the adversarial principle underlying GANs and the Jensen-Shannon divergence, together with some optimality characteristics of the problem. An analysis of the role of the discriminator family via approximation arguments is also provided. In addition, taking a statistical point of view, we study the large sample properties of the estimated distribution and prove in particular a central limit theorem. Some of our results are illustrated with simulated examples.
Type de document :
Pré-publication, Document de travail
Liste complète des métadonnées

Littérature citée [14 références]  Voir  Masquer  Télécharger
Contributeur : Biau Gérard <>
Soumis le : mardi 20 mars 2018 - 09:52:46
Dernière modification le : mardi 19 mars 2019 - 01:23:26
Document(s) archivé(s) le : mardi 11 septembre 2018 - 09:27:49


Fichiers produits par l'(les) auteur(s)


  • HAL Id : hal-01737975, version 1
  • ARXIV : 1803.07819


G. Biau, B. Cadre, M. Sangnier, U. Tanielian. Some Theoretical Properties of GANs. 2018. 〈hal-01737975〉



Consultations de la notice


Téléchargements de fichiers