Matches in SemOpenAlex for { <https://semopenalex.org/work/W3184670645> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W3184670645 endingPage "143" @default.
- W3184670645 startingPage "130" @default.
- W3184670645 abstract "In recent years, even though Stochastic Gradient Descent (SGD) and its variants are well-known for training neural networks, it suffers from limitations such as the lack of theoretical guarantees, vanishing gradients, and excessive sensitivity to input. To overcome these drawbacks, alternating minimization methods have attracted fast-increasing attention recently. As an emerging and open domain, however, several new challenges need to be addressed, including 1) Convergence properties are sensitive to penalty parameters, and 2) Slow theoretical convergence rate. We, therefore, propose a novel monotonous Deep Learning Alternating Minimization (mDLAM) algorithm to deal with these two challenges. Our innovative inequality-constrained formulation infinitely approximates the original problem with non-convex equality constraints, enabling our convergence proof of the proposed mDLAM algorithm regardless of the choice of hyperparameters. Our mDLAM algorithm is shown to achieve a fast linear convergence by the Nesterov acceleration technique. Extensive experiments on multiple benchmark datasets demonstrate the convergence, effectiveness, and efficiency of the proposed mDLAM algorithm." @default.
- W3184670645 created "2021-08-02" @default.
- W3184670645 creator A5000223219 @default.
- W3184670645 creator A5010153102 @default.
- W3184670645 creator A5017662665 @default.
- W3184670645 creator A5059698647 @default.
- W3184670645 date "2022-05-01" @default.
- W3184670645 modified "2023-09-27" @default.
- W3184670645 title "Accelerated Gradient-free Neural Network Training by Multi-convex Alternating Optimization" @default.
- W3184670645 cites W114517082 @default.
- W3184670645 cites W1498436455 @default.
- W3184670645 cites W1968154520 @default.
- W3184670645 cites W1987083649 @default.
- W3184670645 cites W1988720110 @default.
- W3184670645 cites W2000956074 @default.
- W3184670645 cites W2100556411 @default.
- W3184670645 cites W2153959628 @default.
- W3184670645 cites W2594314233 @default.
- W3184670645 cites W3128872986 @default.
- W3184670645 cites W3134578405 @default.
- W3184670645 cites W3136433533 @default.
- W3184670645 doi "https://doi.org/10.1016/j.neucom.2022.02.039" @default.
- W3184670645 hasPublicationYear "2022" @default.
- W3184670645 type Work @default.
- W3184670645 sameAs 3184670645 @default.
- W3184670645 citedByCount "3" @default.
- W3184670645 countsByYear W31846706452020 @default.
- W3184670645 countsByYear W31846706452023 @default.
- W3184670645 crossrefType "journal-article" @default.
- W3184670645 hasAuthorship W3184670645A5000223219 @default.
- W3184670645 hasAuthorship W3184670645A5010153102 @default.
- W3184670645 hasAuthorship W3184670645A5017662665 @default.
- W3184670645 hasAuthorship W3184670645A5059698647 @default.
- W3184670645 hasBestOaLocation W31846706451 @default.
- W3184670645 hasConcept C10494615 @default.
- W3184670645 hasConcept C112680207 @default.
- W3184670645 hasConcept C11413529 @default.
- W3184670645 hasConcept C115680565 @default.
- W3184670645 hasConcept C121332964 @default.
- W3184670645 hasConcept C126255220 @default.
- W3184670645 hasConcept C153258448 @default.
- W3184670645 hasConcept C153294291 @default.
- W3184670645 hasConcept C154945302 @default.
- W3184670645 hasConcept C2524010 @default.
- W3184670645 hasConcept C2777211547 @default.
- W3184670645 hasConcept C33923547 @default.
- W3184670645 hasConcept C41008148 @default.
- W3184670645 hasConcept C50644808 @default.
- W3184670645 hasConceptScore W3184670645C10494615 @default.
- W3184670645 hasConceptScore W3184670645C112680207 @default.
- W3184670645 hasConceptScore W3184670645C11413529 @default.
- W3184670645 hasConceptScore W3184670645C115680565 @default.
- W3184670645 hasConceptScore W3184670645C121332964 @default.
- W3184670645 hasConceptScore W3184670645C126255220 @default.
- W3184670645 hasConceptScore W3184670645C153258448 @default.
- W3184670645 hasConceptScore W3184670645C153294291 @default.
- W3184670645 hasConceptScore W3184670645C154945302 @default.
- W3184670645 hasConceptScore W3184670645C2524010 @default.
- W3184670645 hasConceptScore W3184670645C2777211547 @default.
- W3184670645 hasConceptScore W3184670645C33923547 @default.
- W3184670645 hasConceptScore W3184670645C41008148 @default.
- W3184670645 hasConceptScore W3184670645C50644808 @default.
- W3184670645 hasLocation W31846706451 @default.
- W3184670645 hasLocation W31846706452 @default.
- W3184670645 hasOpenAccess W3184670645 @default.
- W3184670645 hasPrimaryLocation W31846706451 @default.
- W3184670645 hasRelatedWork W1983212821 @default.
- W3184670645 hasRelatedWork W2594900731 @default.
- W3184670645 hasRelatedWork W2811217697 @default.
- W3184670645 hasRelatedWork W3201745389 @default.
- W3184670645 hasRelatedWork W4200512840 @default.
- W3184670645 hasRelatedWork W4254078441 @default.
- W3184670645 hasRelatedWork W4293324523 @default.
- W3184670645 hasRelatedWork W4295266778 @default.
- W3184670645 hasRelatedWork W4318347891 @default.
- W3184670645 hasRelatedWork W4372267246 @default.
- W3184670645 hasVolume "487" @default.
- W3184670645 isParatext "false" @default.
- W3184670645 isRetracted "false" @default.
- W3184670645 magId "3184670645" @default.
- W3184670645 workType "article" @default.