A Neural Tangent Kernel Perspective of GANs - Archive ouverte HAL Access content directly
Poster Communications Year :

A Neural Tangent Kernel Perspective of GANs

Abstract

We propose a novel theoretical framework of analysis for Generative Adversarial Networks (GANs). We reveal a fundamental flaw of previous analyses which, by incorrectly modeling GANs' training scheme, are subject to ill-defined discriminator gradients. We overcome this issue which impedes a principled study of GAN training, solving it within our framework by taking into account the discriminator's architecture. To this end, we leverage the theory of infinite-width neural networks for the discriminator via its Neural Tangent Kernel. We characterize the trained discriminator for a wide range of losses and establish general differentiability properties of the network. From this, we derive new insights about the convergence of the generated distribution, advancing our understanding of GANs' training dynamics. We empirically corroborate these results via an analysis toolkit based on our framework, unveiling intuitions that are consistent with GAN practice.
Fichier principal
Vignette du fichier
gantk2_poster.pdf (913.03 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03716574 , version 1 (07-07-2022)

Licence

Attribution - NonCommercial - ShareAlike

Identifiers

  • HAL Id : hal-03716574 , version 1

Cite

Jean-Yves Franceschi, Emmanuel de Bézenac, Ibrahim Ayed, Mickaël Chen, Sylvain Lamprier, et al.. A Neural Tangent Kernel Perspective of GANs. Thirty-ninth International Conference on Machine Learning, Jul 2022, Baltimore, MD, United States. . ⟨hal-03716574⟩

Relations

46 View
13 Download

Share

Gmail Facebook Twitter LinkedIn More