Variational Bayesian GaN

Jen Tzung Chien, Chun Lin Kuo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

Generative adversarial network (GAN) has been successfully developing as a generative model where the artificial data drawn from the generator are misrecognized as real samples by a discriminator. Although GAN achieves the desirable performance, the challenge is that the mode collapse easily happens in the joint optimization of generator and discriminator. This study copes with this challenge by improving the model regularization by means of representing the weight uncertainty in GAN. A new Bayesian GAN is formulated and implemented to learn a regularized model from diverse data where the strong modes are flattened via the marginalization and the issues of model collapse and gradient vanishing are alleviated. In particular, we present a variational GAN (VGAN) where the encoder, generator and discriminator are jointly estimated according to the variational Bayesian inference. The experiments on image generation over two tasks (MNIST and CeleA) demonstrate the superiority of the proposed VGAN to the variational autoencoder, the standard GAN and the Bayesian GAN based on the sampling method. The learning efficiency and generation performance are evaluated.

Original languageEnglish
Title of host publicationEUSIPCO 2019 - 27th European Signal Processing Conference
PublisherEuropean Signal Processing Conference, EUSIPCO
ISBN (Electronic)9789082797039
DOIs
StatePublished - Sep 2019
Event27th European Signal Processing Conference, EUSIPCO 2019 - A Coruna, Spain
Duration: 2 Sep 20196 Sep 2019

Publication series

NameEuropean Signal Processing Conference
Volume2019-September
ISSN (Print)2219-5491

Conference

Conference27th European Signal Processing Conference, EUSIPCO 2019
CountrySpain
CityA Coruna
Period2/09/196/09/19

Keywords

  • Bayesian learning
  • Computer vision
  • Generative adversarial networks
  • Variational autoencoder

Fingerprint Dive into the research topics of 'Variational Bayesian GaN'. Together they form a unique fingerprint.

Cite this