Show simple item record

dc.contributor.authorKandemir, Melih
dc.date.accessioned2019-04-02T10:58:21Z
dc.date.available2019-04-02T10:58:21Z
dc.date.issued2018-09
dc.identifier.issn0167-8655en_US
dc.identifier.urihttp://hdl.handle.net/10679/6250
dc.identifier.urihttps://www.sciencedirect.com/science/article/pii/S0167865518302861#!
dc.description.abstractWe introduce a Bayesian construction for deep neural networks that is amenable to mean field variational inference that operates solely by closed-form update rules. Hence, it does not require any learning rate to be manually tuned. We show that by this virtue it becomes possible with our model to perform effective deep learning on three setups where conventional neural nets are known to perform suboptimally: i) online learning, ii) learning from small data, and iii) active learning. We compare our approach to earlier Bayesian neural network inference techniques spanning from expectation propagation to gradient-based variational Bayes, as well as deterministic neural nets with various activations functions. We observe our approach to improve on all these alternatives in two mainstream vision benchmarks and two medical data sets: diabetic retinopathy screening and exudate detection from eye fundus images.en_US
dc.language.isoengen_US
dc.publisherElsevieren_US
dc.relation.ispartofPattern Recognition Letters
dc.rightsrestrictedAccess
dc.titleVariational closed-Form deep neural net inferenceen_US
dc.typeArticleen_US
dc.peerreviewedyesen_US
dc.publicationstatusPublisheden_US
dc.contributor.departmentÖzyeğin University
dc.contributor.authorID(ORCID 0000-0001-6293-3656 & YÖK ID 258737) Kandemir, Melih
dc.contributor.ozuauthorKandemir, Melih
dc.identifier.volume112en_US
dc.identifier.startpage145en_US
dc.identifier.endpage151en_US
dc.identifier.wosWOS:000443950800021
dc.identifier.doi10.1016/j.patrec.2018.07.001en_US
dc.subject.keywordsBayesian Neural Networksen_US
dc.subject.keywordsVariational Bayesen_US
dc.subject.keywordsOnline learningen_US
dc.subject.keywordsActive learningen_US
dc.identifier.scopusSCOPUS:2-s2.0-85049560836
dc.contributor.authorMale1
dc.relation.publicationcategoryArticle - International Refereed Journal - Institutional Academic Staff


Files in this item

FilesSizeFormatView

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record


Share this page