Volltext-Downloads (blau) und Frontdoor-Views (grau)
  • search hit 22 of 176
Back to Result List

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.show moreshow less

Export metadata

Additional Services

Share in Twitter Search Google Scholar

Statistics

frontdoor_oas
Metadaten
Author:Ricard Durall Lopez, Avraam Chatzimichailidis, Peter Labus, Janis KeuperORCiDGND
Date of Publication (online):2020/12/17
Pagenumber:9
Language:English
Document Type:Article (unreviewed)
Open Access:Frei zugänglich
Release Date:2021/01/25
Licence (German):License LogoEs gilt das UrhG
Note:
Pre-Print
URL:https://www.researchgate.net/publication/347441697_Combating_Mode_Collapse_in_GAN_training_An_Empirical_Analysis_using_Hessian_Eigenvalues
ArXiv Id:http://arxiv.org/abs/2012.09673