Vis enkel innførsel

dc.contributor.authorKampffmeyer, Michael C.
dc.contributor.authorLøkse, Sigurd
dc.contributor.authorBianchi, Filippo Maria
dc.contributor.authorJenssen, Robert
dc.contributor.authorLivi, Lorenzo
dc.date.accessioned2018-09-18T12:39:18Z
dc.date.available2018-09-18T12:39:18Z
dc.date.issued2017-05-19
dc.description.abstractIn this paper we introduce the deep kernelized autoencoder, a neural network model that allows an explicit approximation of (i) the mapping from an input space to an arbitrary, user-specified kernel space and (ii) the back-projection from such a kernel space to input space. The proposed method is based on traditional autoencoders and is trained through a new unsupervised loss function. During training, we optimize both the reconstruction accuracy of input samples and the alignment between a kernel matrix given as prior and the inner products of the hidden representations computed by the autoencoder. Kernel alignment provides control over the hidden representation learned by the autoen- coder. Experiments have been performed to evaluate both reconstruction and kernel alignment performance. Additionally, we applied our method to emulate kPCA on a denoising task obtaining promising resultsen_US
dc.description.sponsorshipWe gratefully acknowledge the support of NVIDIA Corporation with the donation of the GPU used for this research.en_US
dc.descriptionAccepted manuscript version allowed (see <a href=https://www.springer.com/gp/open-access>policy</a>). <br>Published version available in:<a href=https://link.springer.com/chapter/10.1007/978-3-319-59126-1_35>https://link.springer.com/chapter/10.1007/978-3-319-59126-1_35</a>en_US
dc.identifier.citationKampffmeyer MC, Løkse S, Bianchi FM, Jenssen R, Livi L. Deep kernelized autoencoders. Lecture Notes in Computer Science. 2017;10269 LNCS:419-430 DOI:10.1007/978-3-319-59126-1_35en_US
dc.identifier.cristinIDFRIDAID 1493964
dc.identifier.doi10.1007/978-3-319-59126-1_35
dc.identifier.issn0302-9743
dc.identifier.issn1611-3349
dc.identifier.urihttps://hdl.handle.net/10037/13824
dc.language.isoengen_US
dc.publisherSpringer International Publishingen_US
dc.relation.journalLecture Notes in Computer Science
dc.relation.projectIDNorges forskningsråd: 239844en_US
dc.relation.projectIDNorges forskningsråd: 270738en_US
dc.rights.accessRightsopenAccessen_US
dc.subjectVDP::Technology: 500::Information and communication technology: 550en_US
dc.subjectVDP::Teknologi: 500::Informasjons- og kommunikasjonsteknologi: 550en_US
dc.titleDeep kernelized autoencodersen_US
dc.typePeer revieweden_US
dc.typeBooken_US
dc.typeBokkapittelno
dc.typeBokno
dc.typeChapteren_US


Tilhørende fil(er)

Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel