OPTIMAL ALGORITHMS FOR STOCHASTIC COMPLEMENTARY COMPOSITE MINIMIZATION

Simple item page
dc.contributor.authorD'aspremont, Alexandre
dc.contributor.authorGuzman, Cristobal
dc.contributor.authorLezane, Clement
dc.date.accessioned2025-01-20T16:18:36Z
dc.date.available2025-01-20T16:18:36Z
dc.date.issued2024
dc.description.abstractInspired by regularization techniques in statistics and machine learning, we study complementary composite minimization in the stochastic setting. This problem corresponds to the minimization of the sum of a (weakly) smooth function endowed with a stochastic first -order oracle and a structured uniformly convex (possibly nonsmooth and non -Lipschitz) regularization term. Despite intensive work on closely related settings, prior to our work no complexity bounds for this problem were known. We close this gap by providing novel excess risk bounds, both in expectation and with high probability. Our algorithms are nearly optimal, which we prove via novel lower complexity bounds for this class of problems. We conclude by providing numerical results comparing our methods to the state of the art.
dc.fuente.origenWOS
dc.identifier.doi10.1137/22M1530884
dc.identifier.eissn1095-7189
dc.identifier.issn1052-6234
dc.identifier.urihttps://doi.org/10.1137/22M1530884
dc.identifier.urihttps://repositorio.uc.cl/handle/11534/90655
dc.identifier.wosidWOS:001174991600024
dc.issue.numero1
dc.language.isoen
dc.pagina.final189
dc.pagina.inicio163
dc.revistaSiam journal on optimization
dc.rightsacceso restringido
dc.subjectstochastic convex optimization
dc.subjectregularization
dc.subjectnon-Euclidean composite minimiza- tion
dc.subjectaccelerated first-order methods
dc.titleOPTIMAL ALGORITHMS FOR STOCHASTIC COMPLEMENTARY COMPOSITE MINIMIZATION
dc.typeartículo
dc.volumen34
sipa.indexWOS
sipa.trazabilidadWOS;2025-01-12
Files
Collections
Artículos de revistas