Matches in SemOpenAlex for { <https://semopenalex.org/work/W3131567436> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W3131567436 abstract "Adaptive gradient methods have been shown to outperform SGD in many tasks of training neural networks. However, the acceleration effect is yet to be explained in the non-convex setting since the best convergence rate of adaptive gradient methods is worse than that of SGD in literature. In this paper, we prove that adaptive gradient methods exhibit an O~(T−1/2)-convergence rate for finding first-order stationary points under the strong growth condition, which improves previous best convergence results of adaptive gradient methods and random shuffling SGD by factors of O(T−1/4) and O(T−1/6), respectively. In particular, we study two variants of AdaGrad with random shuffling for finite sum minimization. Our analysis suggests that the combination of random shuffling and adaptive learning rates gives rise to better convergence." @default.
- W3131567436 created "2021-03-01" @default.
- W3131567436 creator A5022574607 @default.
- W3131567436 creator A5055968240 @default.
- W3131567436 creator A5075889094 @default.
- W3131567436 creator A5083207127 @default.
- W3131567436 date "2021-05-04" @default.
- W3131567436 modified "2023-09-23" @default.
- W3131567436 title "Adaptive Gradient Methods Can Be Provably Faster than SGD with Random Shuffling" @default.
- W3131567436 hasPublicationYear "2021" @default.
- W3131567436 type Work @default.
- W3131567436 sameAs 3131567436 @default.
- W3131567436 citedByCount "0" @default.
- W3131567436 crossrefType "journal-article" @default.
- W3131567436 hasAuthorship W3131567436A5022574607 @default.
- W3131567436 hasAuthorship W3131567436A5055968240 @default.
- W3131567436 hasAuthorship W3131567436A5075889094 @default.
- W3131567436 hasAuthorship W3131567436A5083207127 @default.
- W3131567436 hasConcept C105795698 @default.
- W3131567436 hasConcept C11413529 @default.
- W3131567436 hasConcept C126255220 @default.
- W3131567436 hasConcept C134306372 @default.
- W3131567436 hasConcept C147764199 @default.
- W3131567436 hasConcept C154945302 @default.
- W3131567436 hasConcept C162324750 @default.
- W3131567436 hasConcept C167927819 @default.
- W3131567436 hasConcept C189237950 @default.
- W3131567436 hasConcept C26517878 @default.
- W3131567436 hasConcept C2777303404 @default.
- W3131567436 hasConcept C28826006 @default.
- W3131567436 hasConcept C33923547 @default.
- W3131567436 hasConcept C38652104 @default.
- W3131567436 hasConcept C41008148 @default.
- W3131567436 hasConcept C50522688 @default.
- W3131567436 hasConcept C50644808 @default.
- W3131567436 hasConcept C57869625 @default.
- W3131567436 hasConceptScore W3131567436C105795698 @default.
- W3131567436 hasConceptScore W3131567436C11413529 @default.
- W3131567436 hasConceptScore W3131567436C126255220 @default.
- W3131567436 hasConceptScore W3131567436C134306372 @default.
- W3131567436 hasConceptScore W3131567436C147764199 @default.
- W3131567436 hasConceptScore W3131567436C154945302 @default.
- W3131567436 hasConceptScore W3131567436C162324750 @default.
- W3131567436 hasConceptScore W3131567436C167927819 @default.
- W3131567436 hasConceptScore W3131567436C189237950 @default.
- W3131567436 hasConceptScore W3131567436C26517878 @default.
- W3131567436 hasConceptScore W3131567436C2777303404 @default.
- W3131567436 hasConceptScore W3131567436C28826006 @default.
- W3131567436 hasConceptScore W3131567436C33923547 @default.
- W3131567436 hasConceptScore W3131567436C38652104 @default.
- W3131567436 hasConceptScore W3131567436C41008148 @default.
- W3131567436 hasConceptScore W3131567436C50522688 @default.
- W3131567436 hasConceptScore W3131567436C50644808 @default.
- W3131567436 hasConceptScore W3131567436C57869625 @default.
- W3131567436 hasOpenAccess W3131567436 @default.
- W3131567436 hasRelatedWork W1482718813 @default.
- W3131567436 hasRelatedWork W1487944798 @default.
- W3131567436 hasRelatedWork W1497593879 @default.
- W3131567436 hasRelatedWork W1521926339 @default.
- W3131567436 hasRelatedWork W2003702327 @default.
- W3131567436 hasRelatedWork W2018625446 @default.
- W3131567436 hasRelatedWork W2075524109 @default.
- W3131567436 hasRelatedWork W2127063521 @default.
- W3131567436 hasRelatedWork W2158198268 @default.
- W3131567436 hasRelatedWork W2163457743 @default.
- W3131567436 hasRelatedWork W2358176524 @default.
- W3131567436 hasRelatedWork W2374559649 @default.
- W3131567436 hasRelatedWork W2388411337 @default.
- W3131567436 hasRelatedWork W2393449303 @default.
- W3131567436 hasRelatedWork W2546804429 @default.
- W3131567436 hasRelatedWork W2745847213 @default.
- W3131567436 hasRelatedWork W2786899637 @default.
- W3131567436 hasRelatedWork W2885751302 @default.
- W3131567436 hasRelatedWork W3028175200 @default.
- W3131567436 hasRelatedWork W1902982579 @default.
- W3131567436 isParatext "false" @default.
- W3131567436 isRetracted "false" @default.
- W3131567436 magId "3131567436" @default.
- W3131567436 workType "article" @default.