Show simple item record

dc.contributor.authorIvanova, Anastasiya
dc.contributor.authorGrishchenko, Dmitry
dc.contributor.authorGasnikov, Alexander
dc.contributor.authorShulgin, Egor
dc.date.accessioned2019-12-23T06:33:12Z
dc.date.available2019-12-23T06:33:12Z
dc.date.issued2019-11-25
dc.identifier.urihttp://hdl.handle.net/10754/660741
dc.description.abstractIn 2015 there appears a universal framework Catalyst that allows to accelerate almost arbitrary non-accelerated deterministic and randomized algorithms for smooth convex optimization problems Lin et al. (2015). This technique finds a lot of applications in Machine Learning due to the possibility to deal with sum-type target functions. The significant part of the Catalyst approach is accelerated proximal outer gradient method. This method used as an envelope for non-accelerated inner algorithm for the regularized auxiliary problem. One of the main practical problem of this approach is the selection of this regularization parameter. There exists a nice theory for that at Lin et al. (2018), but this theory required prior knowledge about the smoothness constant of the target function. In this paper, we propose an adaptive variant of Catalyst that doesn't require such information. In combination with the adaptive inner non-accelerated algorithm, we propose accelerated variants of well-known methods: steepest descent, adaptive coordinate descent, alternating minimization.
dc.description.sponsorshipWe would like to thank Pavel Dvurechensky (WIAS, Berlin) and Peter Richtarik (KAUST) for useful remarks.
dc.publisherarXiv
dc.relation.urlhttps://arxiv.org/pdf/1911.11271
dc.rightsArchived with thanks to arXiv
dc.titleAdaptive Catalyst for smooth convex optimization
dc.typePreprint
dc.eprint.versionPre-print
dc.contributor.institutionMoscow Institute of Physics and Technology, Moscow, Russia
dc.contributor.institutionNational Research University Higher School of Economics, Moscow, Russia
dc.contributor.institutionUniversit´e Grenoble Alpes, Grenoble, France
dc.contributor.institutionInstitute for Information Transmission Problems, Moscow, Russia
dc.contributor.institutionCaucasus Mathematical Center, Adyghe State University, Russia
dc.identifier.arxivid1911.11271
refterms.dateFOA2019-12-23T06:33:32Z


Files in this item

Thumbnail
Name:
Preprintfile1.pdf
Size:
548.9Kb
Format:
PDF
Description:
Pre-print

This item appears in the following Collection(s)

Show simple item record