Accelerated Bregman proximal gradient methods for relatively smooth convex optimization
dc.contributor.author | Hanzely, Filip | |
dc.contributor.author | Richtarik, Peter | |
dc.contributor.author | Xiao, Lin | |
dc.date.accessioned | 2021-04-21T07:56:17Z | |
dc.date.available | 2019-05-29T07:28:01Z | |
dc.date.available | 2021-04-21T07:56:17Z | |
dc.date.issued | 2021-04-07 | |
dc.date.submitted | 2020-04-24 | |
dc.identifier.citation | Hanzely, F., Richtárik, P., & Xiao, L. (2021). Accelerated Bregman proximal gradient methods for relatively smooth convex optimization. Computational Optimization and Applications. doi:10.1007/s10589-021-00273-8 | |
dc.identifier.issn | 1573-2894 | |
dc.identifier.issn | 0926-6003 | |
dc.identifier.doi | 10.1007/s10589-021-00273-8 | |
dc.identifier.uri | http://hdl.handle.net/10754/653113 | |
dc.description.abstract | We consider the problem of minimizing the sum of two convex functions: one is differentiable and relatively smooth with respect to a reference convex function, and the other can be nondifferentiable but simple to optimize. We investigate a triangle scaling property of the Bregman distance generated by the reference convex function and present accelerated Bregman proximal gradient (ABPG) methods that attain an O(k-γ) convergence rate, where γ∈ (0 , 2] is the triangle scaling exponent (TSE) of the Bregman distance. For the Euclidean distance, we have γ= 2 and recover the convergence rate of Nesterov’s accelerated gradient methods. For non-Euclidean Bregman distances, the TSE can be much smaller (say γ≤ 1), but we show that a relaxed definition of intrinsic TSE is always equal to 2. We exploit the intrinsic TSE to develop adaptive ABPG methods that converge much faster in practice. Although theoretical guarantees on a fast convergence rate seem to be out of reach in general, our methods obtain empirical O(k- 2) rates in numerical experiments on several applications and provide posterior numerical certificates for the fast rates. | |
dc.description.sponsorship | We thank Haihao Lu, Robert Freund and Yurii Nesterov for helpful conversations. We are also grateful to the anonymous referees whose comments helped improving the clarity of the paper. Peter Richtárik acknowledges the support of the KAUST Baseline Research Funding Scheme. | |
dc.publisher | Springer Science and Business Media LLC | |
dc.relation.url | http://link.springer.com/10.1007/s10589-021-00273-8 | |
dc.rights | Archived with thanks to Computational Optimization and Applications | |
dc.title | Accelerated Bregman proximal gradient methods for relatively smooth convex optimization | |
dc.type | Article | |
dc.contributor.department | Applied Mathematics and Computational Science Program | |
dc.contributor.department | Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division | |
dc.contributor.department | Computer Science Program | |
dc.identifier.journal | Computational Optimization and Applications | |
dc.rights.embargodate | 2022-04-07 | |
dc.eprint.version | Post-print | |
dc.contributor.institution | Toyota Technological Institute at Chicago (TTIC), Chicago, Illinois, USA | |
dc.contributor.institution | Moscow Institute of Physics and Technology, Dolgoprudny, Russia | |
dc.contributor.institution | Microsoft Research, Redmond, WA, USA | |
dc.identifier.arxivid | 1808.03045 | |
kaust.person | Hanzely, Filip | |
kaust.person | Richtarik, Peter | |
dc.version | v1 | |
dc.identifier.eid | 2-s2.0-85103928357 | |
refterms.dateFOA | 2019-05-29T07:28:12Z | |
kaust.acknowledged.supportUnit | KAUST Baseline Research Funding Scheme |
Files in this item
This item appears in the following Collection(s)
-
Articles
-
Applied Mathematics and Computational Science Program
For more information visit: https://cemse.kaust.edu.sa/amcs -
Computer Science Program
For more information visit: https://cemse.kaust.edu.sa/cs -
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
For more information visit: https://cemse.kaust.edu.sa/