Volltext-Downloads (blau) und Frontdoor-Views (grau)

Combating Mode Collapse in GAN training: An Empirical Analysis using Hessian Eigenvalues

  • Generative adversarial networks (GANs) provide state-of-the-art results in image generation. However, despite being so powerful, they still remain very challenging to train. This is in particular caused by their highly non-convex optimization space leading to a number of instabilities. Among them, mode collapse stands out as one of the most daunting ones. This undesirable event occurs when theGenerative adversarial networks (GANs) provide state-of-the-art results in image generation. However, despite being so powerful, they still remain very challenging to train. This is in particular caused by their highly non-convex optimization space leading to a number of instabilities. Among them, mode collapse stands out as one of the most daunting ones. This undesirable event occurs when the model can only fit a few modes of the data distribution, while ignoring the majority of them. In this work, we combat mode collapse using second-order gradient information. To do so, we analyse the loss surface through its Hessian eigenvalues, and show that mode collapse is related to the convergence towards sharp minima. In particular, we observe how the eigenvalues of the G are directly correlated with the occurrence of mode collapse. Finally, motivated by these findings, we design a new optimization algorithm called nudged-Adam (NuGAN) that uses spectral information to overcome mode collapse, leading to empirically more stable convergence properties.…zeige mehrzeige weniger

Metadaten exportieren

Weitere Dienste

Suche bei Google Scholar

Statistik

frontdoor_oas
Metadaten
Dokumentart:Preprint
Zitierlink: https://opus.hs-offenburg.de/4623
Bibliografische Angaben
Titel (Englisch):Combating Mode Collapse in GAN training: An Empirical Analysis using Hessian Eigenvalues
Verfasserangaben:Ricard Durall Lopez, Avraam Chatzimichailidis, Peter Labus, Janis KeuperStaff MemberORCiDGND
Datum der Veröffentlichung (online):17.12.2020
Seitenanzahl:9
URL:https://www.researchgate.net/publication/347441697
Sprache:Englisch
Inhaltliche Informationen
Fakultäten / Einrichtungen:Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019)
Forschung:IMLA - Institute for Machine Learning and Analytics
Sammlungen der Hochschule Offenburg:Bibliografie
Formale Angaben
Open-Access-Status: Open Access 
Lizenz (Deutsch):License LogoUrheberrechtlich geschützt
ArXiv-Id:http://arxiv.org/abs/2012.09673