Matches in SemOpenAlex for { <https://semopenalex.org/work/W3099502502> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W3099502502 endingPage "844" @default.
- W3099502502 startingPage "833" @default.
- W3099502502 abstract "In this paper, we introduce a simplified and unified method for finite-sum convex optimization, named emph{Variance Reduction via Accelerated Dual Averaging (VRADA)}. In both general convex and strongly convex settings, VRADA can attain an $Obig(frac{1}{n}big)$-accurate solution in $O(nloglog n)$ number of stochastic gradient evaluations which improves the best-known result $O(nlog n)$, where $n$ is the number of samples. Meanwhile, VRADA matches the lower bound of the general convex setting up to a $loglog n$ factor and matches the lower bounds in both regimes $nle Theta(kappa)$ and $ngg kappa$ of the strongly convex setting, where $kappa$ denotes the condition number. Besides improving the best-known results and matching all the above lower bounds simultaneously, VRADA has more unified and simplified algorithmic implementation and convergence analysis for both the general convex and strongly convex settings. The underlying novel approaches such as the novel initialization strategy in VRADA may be of independent interest. Through experiments on real datasets, we show the good performance of VRADA over existing methods for large-scale machine learning problems." @default.
- W3099502502 created "2020-11-23" @default.
- W3099502502 creator A5029760958 @default.
- W3099502502 creator A5033421780 @default.
- W3099502502 creator A5049491250 @default.
- W3099502502 date "2020-01-01" @default.
- W3099502502 modified "2023-09-24" @default.
- W3099502502 title "Variance Reduction via Accelerated Dual Averaging for Finite-Sum Optimization" @default.
- W3099502502 hasPublicationYear "2020" @default.
- W3099502502 type Work @default.
- W3099502502 sameAs 3099502502 @default.
- W3099502502 citedByCount "5" @default.
- W3099502502 countsByYear W30995025022021 @default.
- W3099502502 crossrefType "proceedings-article" @default.
- W3099502502 hasAuthorship W3099502502A5029760958 @default.
- W3099502502 hasAuthorship W3099502502A5033421780 @default.
- W3099502502 hasAuthorship W3099502502A5049491250 @default.
- W3099502502 hasConcept C105795698 @default.
- W3099502502 hasConcept C111335779 @default.
- W3099502502 hasConcept C112680207 @default.
- W3099502502 hasConcept C114466953 @default.
- W3099502502 hasConcept C114614502 @default.
- W3099502502 hasConcept C118615104 @default.
- W3099502502 hasConcept C126255220 @default.
- W3099502502 hasConcept C134306372 @default.
- W3099502502 hasConcept C145446738 @default.
- W3099502502 hasConcept C157972887 @default.
- W3099502502 hasConcept C162324750 @default.
- W3099502502 hasConcept C165064840 @default.
- W3099502502 hasConcept C19499675 @default.
- W3099502502 hasConcept C199360897 @default.
- W3099502502 hasConcept C2524010 @default.
- W3099502502 hasConcept C2777303404 @default.
- W3099502502 hasConcept C33923547 @default.
- W3099502502 hasConcept C41008148 @default.
- W3099502502 hasConcept C50522688 @default.
- W3099502502 hasConcept C62644790 @default.
- W3099502502 hasConcept C63553672 @default.
- W3099502502 hasConcept C77553402 @default.
- W3099502502 hasConceptScore W3099502502C105795698 @default.
- W3099502502 hasConceptScore W3099502502C111335779 @default.
- W3099502502 hasConceptScore W3099502502C112680207 @default.
- W3099502502 hasConceptScore W3099502502C114466953 @default.
- W3099502502 hasConceptScore W3099502502C114614502 @default.
- W3099502502 hasConceptScore W3099502502C118615104 @default.
- W3099502502 hasConceptScore W3099502502C126255220 @default.
- W3099502502 hasConceptScore W3099502502C134306372 @default.
- W3099502502 hasConceptScore W3099502502C145446738 @default.
- W3099502502 hasConceptScore W3099502502C157972887 @default.
- W3099502502 hasConceptScore W3099502502C162324750 @default.
- W3099502502 hasConceptScore W3099502502C165064840 @default.
- W3099502502 hasConceptScore W3099502502C19499675 @default.
- W3099502502 hasConceptScore W3099502502C199360897 @default.
- W3099502502 hasConceptScore W3099502502C2524010 @default.
- W3099502502 hasConceptScore W3099502502C2777303404 @default.
- W3099502502 hasConceptScore W3099502502C33923547 @default.
- W3099502502 hasConceptScore W3099502502C41008148 @default.
- W3099502502 hasConceptScore W3099502502C50522688 @default.
- W3099502502 hasConceptScore W3099502502C62644790 @default.
- W3099502502 hasConceptScore W3099502502C63553672 @default.
- W3099502502 hasConceptScore W3099502502C77553402 @default.
- W3099502502 hasLocation W30995025021 @default.
- W3099502502 hasOpenAccess W3099502502 @default.
- W3099502502 hasPrimaryLocation W30995025021 @default.
- W3099502502 hasRelatedWork W2017180575 @default.
- W3099502502 hasRelatedWork W2107438106 @default.
- W3099502502 hasRelatedWork W2149479912 @default.
- W3099502502 hasRelatedWork W2159723742 @default.
- W3099502502 hasRelatedWork W2171687261 @default.
- W3099502502 hasRelatedWork W2342780126 @default.
- W3099502502 hasRelatedWork W2402721272 @default.
- W3099502502 hasRelatedWork W2417107316 @default.
- W3099502502 hasRelatedWork W2471449337 @default.
- W3099502502 hasRelatedWork W2574803746 @default.
- W3099502502 hasRelatedWork W2621572195 @default.
- W3099502502 hasRelatedWork W2914000446 @default.
- W3099502502 hasRelatedWork W2938923584 @default.
- W3099502502 hasRelatedWork W2946943162 @default.
- W3099502502 hasRelatedWork W2953382111 @default.
- W3099502502 hasRelatedWork W3046366290 @default.
- W3099502502 hasRelatedWork W3086294021 @default.
- W3099502502 hasRelatedWork W3131753936 @default.
- W3099502502 hasRelatedWork W3135626733 @default.
- W3099502502 hasRelatedWork W3212554751 @default.
- W3099502502 hasVolume "33" @default.
- W3099502502 isParatext "false" @default.
- W3099502502 isRetracted "false" @default.
- W3099502502 magId "3099502502" @default.
- W3099502502 workType "article" @default.