Skip navigation
Skip navigation

Step size-adapted online support vector learning

Karatzoglou, Alexandros; Vishwanathan, S; Schraudolph, Nicol; Smola, Alexander

Description

We present an online Support Vector Machine (SVM) that uses Stochastic Meta-Descent (SMD) to adapt its step size automatically. We formulate the online learning problem as a stochastic gradient descent in Reproducing Kernel Hubert Space (RKHS) and translate SMD to the nonparametric setting, where its gradient trace parameter is no longer a coefficient vector but an element of the RKHS. We derive efficient updates that allow us to perform the step size adaptation in linear time. We apply the...[Show more]

dc.contributor.authorKaratzoglou, Alexandros
dc.contributor.authorVishwanathan, S
dc.contributor.authorSchraudolph, Nicol
dc.contributor.authorSmola, Alexander
dc.coverage.spatialSydney Australia
dc.date.accessioned2015-12-13T23:00:56Z
dc.date.createdAugust 28 2005
dc.identifier.isbn0780392434
dc.identifier.urihttp://hdl.handle.net/1885/84361
dc.description.abstractWe present an online Support Vector Machine (SVM) that uses Stochastic Meta-Descent (SMD) to adapt its step size automatically. We formulate the online learning problem as a stochastic gradient descent in Reproducing Kernel Hubert Space (RKHS) and translate SMD to the nonparametric setting, where its gradient trace parameter is no longer a coefficient vector but an element of the RKHS. We derive efficient updates that allow us to perform the step size adaptation in linear time. We apply the online SVM framework to a variety of loss functions and in particular show how to achieve efficient online multiclass classification. Experimental evidence suggests that our algorithm outperforms existing methods.
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE Inc)
dc.relation.ispartofseriesInternational Symposium on Signal Processing and Its Applications (ISSPA 2005)
dc.sourceProceedings of the Eighth International Symposium on Signal Processing and Its Applications, 2005
dc.source.urihttp://www.elec.uow.edu.au/isspa2005/
dc.subjectKeywords: Online multiclass classification; Reproducing Kernel Hubert Space (RKHS); Support Vector Machine (SVM); Algorithms; Classification (of information); Metadata; Online systems; Stochastic control systems; Vector quantization; Learning systems
dc.titleStep size-adapted online support vector learning
dc.typeConference paper
local.description.notesImported from ARIES
local.description.refereedYes
dc.date.issued2005
local.identifier.absfor080109 - Pattern Recognition and Data Mining
local.identifier.ariespublicationMigratedxPub12640
local.type.statusPublished Version
local.contributor.affiliationKaratzoglou, Alexandros, Vienna University of Technology
local.contributor.affiliationVishwanathan, S, College of Engineering and Computer Science, ANU
local.contributor.affiliationSchraudolph, Nicol, College of Engineering and Computer Science, ANU
local.contributor.affiliationSmola, Alexander, College of Engineering and Computer Science, ANU
local.description.embargo2037-12-31
local.bibliographicCitation.startpage823
local.bibliographicCitation.lastpage826
local.identifier.doi10.1109/ISSPA.2005.1581065
dc.date.updated2015-12-12T07:36:56Z
local.identifier.scopusID2-s2.0-33847145394
CollectionsANU Research Publications

Download

File Description SizeFormat Image
01_Karatzoglou_Step_size-adapted_online_2005.pdf282.09 kBAdobe PDF    Request a copy


Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.

Updated:  17 November 2022/ Responsible Officer:  University Librarian/ Page Contact:  Library Systems & Web Coordinator