Show simple item record

dc.contributor.authorYu, T.T.
dc.contributor.authorLiu, X.W.
dc.contributor.authorDai, Y.H.
dc.contributor.authorSun, Jie
dc.date.accessioned2023-04-16T09:19:25Z
dc.date.available2023-04-16T09:19:25Z
dc.date.issued2022
dc.identifier.citationYu, T.T. and Liu, X.W. and Dai, Y.H. and Sun, J. 2022. A Mini-Batch Proximal Stochastic Recursive Gradient Algorithm with Diagonal Barzilai–Borwein Stepsize. Journal of the Operations Research Society of China.11: pp. 277-307.
dc.identifier.urihttp://hdl.handle.net/20.500.11937/91425
dc.identifier.doi10.1007/s40305-022-00436-2
dc.description.abstract

Many machine learning problems can be formulated as minimizing the sum of a function and a non-smooth regularization term. Proximal stochastic gradient methods are popular for solving such composite optimization problems. We propose a mini-batch proximal stochastic recursive gradient algorithm SRG-DBB, which incorporates the diagonal Barzilai–Borwein (DBB) stepsize strategy to capture the local geometry of the problem. The linear convergence and complexity of SRG-DBB are analyzed for strongly convex functions. We further establish the linear convergence of SRG-DBB under the non-strong convexity condition. Moreover, it is proved that SRG-DBB converges sublinearly in the convex case. Numerical experiments on standard data sets indicate that the performance of SRG-DBB is better than or comparable to the proximal stochastic recursive gradient algorithm with best-tuned scalar stepsizes or BB stepsizes. Furthermore, SRG-DBB is superior to some advanced mini-batch proximal stochastic gradient methods.

dc.titleA Mini-Batch Proximal Stochastic Recursive Gradient Algorithm with Diagonal Barzilai–Borwein Stepsize
dc.typeJournal Article
dcterms.source.volume11
dcterms.source.startPage277
dcterms.source.endPage307
dcterms.source.issn2194-668X
dcterms.source.titleJournal of the Operations Research Society of China
dc.date.updated2023-04-16T09:19:24Z
curtin.note

This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1007/s40305-022-00436-2.

curtin.departmentSchool of Elec Eng, Comp and Math Sci (EECMS)
curtin.accessStatusOpen access
curtin.facultyFaculty of Science and Engineering
curtin.contributor.orcidSun, Jie [0000-0001-5611-1672]
curtin.contributor.researcheridSun, Jie [B-7926-2016] [G-3522-2010]
dcterms.source.eissn2194-6698
curtin.contributor.scopusauthoridSun, Jie [16312754600] [57190212842]
curtin.repositoryagreementV3


Files in this item

Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record