Matches in SemOpenAlex for { <https://semopenalex.org/work/W3130060520> ?p ?o ?g. }
- W3130060520 endingPage "2262" @default.
- W3130060520 startingPage "2230" @default.
- W3130060520 abstract "We introduce two algorithms for nonconvex regularized finite sum minimization, where typical Lipschitz differentiability assumptions are relaxed to the notion of relative smoothness. The first one is a Bregman extension of Finito/MISO, studied for fully nonconvex problems when the sampling is random, or under convexity of the nonsmooth term when it is essentially cyclic. The second algorithm is a low-memory variant, in the spirit of SVRG and SARAH, that also allows for fully nonconvex formulations. Our analysis is made remarkably simple by employing a Bregman Moreau envelope as Lyapunov function. In the randomized case, linear convergence is established when the cost function is strongly convex, yet with no convexity requirements on the individual functions in the sum. For the essentially cyclic and low-memory variants, global and linear convergence results are established when the cost function satisfies the Kurdyka-L ojasiewicz property." @default.
- W3130060520 created "2021-03-01" @default.
- W3130060520 creator A5010064441 @default.
- W3130060520 creator A5031380765 @default.
- W3130060520 creator A5060005608 @default.
- W3130060520 creator A5063926801 @default.
- W3130060520 date "2022-09-01" @default.
- W3130060520 modified "2023-10-15" @default.
- W3130060520 title "Bregman Finito/MISO for Nonconvex Regularized Finite Sum Minimization without Lipschitz Gradient Continuity" @default.
- W3130060520 cites W1499137793 @default.
- W3130060520 cites W1540764732 @default.
- W3130060520 cites W1577337550 @default.
- W3130060520 cites W1967138577 @default.
- W3130060520 cites W1971540397 @default.
- W3130060520 cites W1971942712 @default.
- W3130060520 cites W1988795359 @default.
- W3130060520 cites W1992208280 @default.
- W3130060520 cites W1998880679 @default.
- W3130060520 cites W2015911499 @default.
- W3130060520 cites W2016384870 @default.
- W3130060520 cites W2017680740 @default.
- W3130060520 cites W2017938640 @default.
- W3130060520 cites W2021361347 @default.
- W3130060520 cites W2022164193 @default.
- W3130060520 cites W2022541060 @default.
- W3130060520 cites W2023901033 @default.
- W3130060520 cites W2027982384 @default.
- W3130060520 cites W2039050532 @default.
- W3130060520 cites W2047152541 @default.
- W3130060520 cites W2073750241 @default.
- W3130060520 cites W2078853581 @default.
- W3130060520 cites W2090376964 @default.
- W3130060520 cites W2092759407 @default.
- W3130060520 cites W2095984592 @default.
- W3130060520 cites W2120717492 @default.
- W3130060520 cites W2129732816 @default.
- W3130060520 cites W2155583876 @default.
- W3130060520 cites W2164224215 @default.
- W3130060520 cites W2331639718 @default.
- W3130060520 cites W2535496140 @default.
- W3130060520 cites W2551153860 @default.
- W3130060520 cites W2552658666 @default.
- W3130060520 cites W2698354080 @default.
- W3130060520 cites W2763081248 @default.
- W3130060520 cites W2800693632 @default.
- W3130060520 cites W2926523766 @default.
- W3130060520 cites W2953690514 @default.
- W3130060520 cites W2962913067 @default.
- W3130060520 cites W2963156201 @default.
- W3130060520 cites W2963648037 @default.
- W3130060520 cites W2963681631 @default.
- W3130060520 cites W2963967579 @default.
- W3130060520 cites W2964211254 @default.
- W3130060520 cites W2964317091 @default.
- W3130060520 cites W2990601391 @default.
- W3130060520 cites W3000307639 @default.
- W3130060520 cites W3021534257 @default.
- W3130060520 cites W3037678684 @default.
- W3130060520 cites W3102206315 @default.
- W3130060520 cites W3113425034 @default.
- W3130060520 cites W3168993225 @default.
- W3130060520 cites W4249667877 @default.
- W3130060520 doi "https://doi.org/10.1137/21m140376x" @default.
- W3130060520 hasPublicationYear "2022" @default.
- W3130060520 type Work @default.
- W3130060520 sameAs 3130060520 @default.
- W3130060520 citedByCount "1" @default.
- W3130060520 countsByYear W31300605202022 @default.
- W3130060520 crossrefType "journal-article" @default.
- W3130060520 hasAuthorship W3130060520A5010064441 @default.
- W3130060520 hasAuthorship W3130060520A5031380765 @default.
- W3130060520 hasAuthorship W3130060520A5060005608 @default.
- W3130060520 hasAuthorship W3130060520A5063926801 @default.
- W3130060520 hasBestOaLocation W31300605202 @default.
- W3130060520 hasConcept C126255220 @default.
- W3130060520 hasConcept C134306372 @default.
- W3130060520 hasConcept C147764199 @default.
- W3130060520 hasConcept C22324862 @default.
- W3130060520 hasConcept C28826006 @default.
- W3130060520 hasConcept C33923547 @default.
- W3130060520 hasConceptScore W3130060520C126255220 @default.
- W3130060520 hasConceptScore W3130060520C134306372 @default.
- W3130060520 hasConceptScore W3130060520C147764199 @default.
- W3130060520 hasConceptScore W3130060520C22324862 @default.
- W3130060520 hasConceptScore W3130060520C28826006 @default.
- W3130060520 hasConceptScore W3130060520C33923547 @default.
- W3130060520 hasFunder F4320321390 @default.
- W3130060520 hasFunder F4320321730 @default.
- W3130060520 hasFunder F4320322308 @default.
- W3130060520 hasFunder F4320334764 @default.
- W3130060520 hasIssue "3" @default.
- W3130060520 hasLocation W31300605201 @default.
- W3130060520 hasLocation W31300605202 @default.
- W3130060520 hasLocation W31300605203 @default.
- W3130060520 hasLocation W31300605204 @default.
- W3130060520 hasLocation W31300605205 @default.
- W3130060520 hasOpenAccess W3130060520 @default.
- W3130060520 hasPrimaryLocation W31300605201 @default.