Dirichlet Variational Autoencoder

Cited 2 time in webofscience Cited 0 time in scopus
  • Hit : 69
  • Download : 0
DC FieldValueLanguage
dc.contributor.authorJoo, Weonyoungko
dc.contributor.authorLee, Wonsungko
dc.contributor.authorPark, Sungraeko
dc.contributor.authorMoon, Il-Chulko
dc.date.accessioned2021-03-26T01:34:18Z-
dc.date.available2021-03-26T01:34:18Z-
dc.date.created2020-08-19-
dc.date.issued2020-11-
dc.identifier.citationPATTERN RECOGNITION, v.107-
dc.identifier.issn0031-3203-
dc.identifier.urihttp://hdl.handle.net/10203/281842-
dc.description.abstractThis paper proposes Dirichlet Variational Autoencoder (DirVAE) using a Dirichlet prior. To infer the parameters of DirVAE, we utilize the stochastic gradient method by approximating the inverse cumulative distribution function of the Gamma distribution, which is a component of the Dirichlet distribution. This approximation on a new prior led an investigation on the component collapsing, and DirVAE revealed that the component collapsing originates from two problem sources: decoder weight collapsing and latent value collapsing. The experimental results show that 1) DirVAE generates the result with the best log-likelihood compared to the baselines; 2) DirVAE produces more interpretable latent values with no collapsing issues which the baselines suffer from; 3) the latent representation from DirVAE achieves the best classification accuracy in the (semi-)supervised classification tasks on MNIST, OMNIGLOT, COIL-20, SVHN, and CIFAR-10 compared to the baseline VAEs; and 4) the DirVAE augmented topic models show better performances in most cases.-
dc.languageEnglish-
dc.publisherELSEVIER SCI LTD-
dc.titleDirichlet Variational Autoencoder-
dc.typeArticle-
dc.identifier.wosid000552866000055-
dc.identifier.scopusid2-s2.0-85087402885-
dc.type.rimsART-
dc.citation.volume107-
dc.citation.publicationnamePATTERN RECOGNITION-
dc.identifier.doi10.1016/j.patcog.2020.107514-
dc.contributor.localauthorMoon, Il-Chul-
dc.contributor.nonIdAuthorLee, Wonsung-
dc.contributor.nonIdAuthorPark, Sungrae-
dc.description.isOpenAccessN-
dc.type.journalArticleArticle-
dc.subject.keywordAuthorRepresentation learning-
dc.subject.keywordAuthorVariational autoencoder-
dc.subject.keywordAuthorDeep generative model-
dc.subject.keywordAuthorMulti-modal latent representation-
dc.subject.keywordAuthorComponent collapse-
Appears in Collection
IE-Journal Papers(저널논문)
Files in This Item
There are no files associated with this item.
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 2 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0