Afficher la notice abrégée

Probabilistic Low-Rank Matrix Completion with Adaptive Spectral Regularization Algorithms

hal.structure.identifierAdvanced Learning Evolutionary Algorithms [ALEA]
hal.structure.identifierInstitut de Mathématiques de Bordeaux [IMB]
hal.structure.identifierQuality control and dynamic reliability [CQFD]
dc.contributor.authorTODESCHINI, Adrien
hal.structure.identifierDepartment of Statistics [Oxford]
dc.contributor.authorCARON, Francois
hal.structure.identifierInstitut de Mathématiques de Bordeaux [IMB]
hal.structure.identifierQuality control and dynamic reliability [CQFD]
dc.contributor.authorCHAVENT, Marie
dc.contributor.editorBurges
dc.contributor.editorC. and Bottou
dc.contributor.editorL. and Welling
dc.contributor.editorM. and Ghahramani
dc.contributor.editorZ. and Weinberger
dc.contributor.editorK.
dc.date.created2013-06
dc.date.issued2013-12
dc.date.conference2013-12
dc.description.abstractEnWe propose a novel class of algorithms for low rank matrix completion. Our approach builds on novel penalty functions on the singular values of the low rank matrix. By exploiting a mixture model representation of this penalty, we show that a suitably chosen set of latent variables enables to derive an Expectation-Maximization algorithm to obtain a Maximum A Posteriori estimate of the completed low rank matrix. The resulting algorithm is an iterative soft-thresholded algorithm which iteratively adapts the shrinkage coefficients associated to the singular values. The algorithm is simple to implement and can scale to large matrices. We provide numerical comparisons between our approach and recent alternatives showing the interest of the proposed approach for low rank matrix completion.
dc.language.isoen
dc.publisherCurran Associates, Inc.
dc.title.enProbabilistic Low-Rank Matrix Completion with Adaptive Spectral Regularization Algorithms
dc.typeCommunication dans un congrès
dc.subject.halStatistiques [stat]/Machine Learning [stat.ML]
dc.subject.halStatistiques [stat]/Méthodologie [stat.ME]
dc.subject.halStatistiques [stat]/Calcul [stat.CO]
dc.subject.halStatistiques [stat]/Applications [stat.AP]
bordeaux.page845-853
bordeaux.volume26
bordeaux.conference.titleNIPS - The Neural Information Processing Systems Conference
bordeaux.countryUS
bordeaux.conference.citySouth Lake Tahoe
bordeaux.peerReviewedoui
hal.identifierhal-01025508
hal.version1
hal.invitednon
hal.proceedingsoui
hal.conference.organizerThe Neural Information Processing Systems (NIPS) Foundation, Inc.
hal.conference.end2013-12
hal.popularnon
hal.audienceInternationale
hal.origin.linkhttps://hal.archives-ouvertes.fr//hal-01025508v1
bordeaux.COinSctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.date=2013-12&rft.volume=26&rft.spage=845-853&rft.epage=845-853&rft.au=TODESCHINI,%20Adrien&CARON,%20Francois&CHAVENT,%20Marie&rft.genre=unknown


Fichier(s) constituant ce document

FichiersTailleFormatVue

Il n'y a pas de fichiers associés à ce document.

Ce document figure dans la(les) collection(s) suivante(s)

Afficher la notice abrégée