Matches in SemOpenAlex for { <https://semopenalex.org/work/W3126996176> ?p ?o ?g. }
Showing items 1 to 88 of
88
with 100 items per page.
- W3126996176 endingPage "23024" @default.
- W3126996176 startingPage "23007" @default.
- W3126996176 abstract "In the field of neural network, stochastic gradient descent is often employed as an effective method of accelerating the result's convergence. Generating the new gradient from the past gradient is a common method adopted by many existing optimization algorithms. Since the past gradient is not computed based on the most updated stochastic gradient descent state, it can introduce a deviation to the new gradient computation, negatively impacting its rate of convergence. To resolve this problem, we propose an algorithm that quantifies this deviation based on the angle between the past and the current gradients, which is then applied to calibrate these two gradients, generating a more accurate new gradient. To demonstrate the broad applicability of the algorithm, the proposed method is implemented into a neural network and a logistic regression classifier which are evaluated on the datasets MNIST and NSL-KDD, respectively. An in-depth analysis is performed to compare our algorithm with nine optimization algorithms in two experiments, demonstrating the advantages in the cost and the error rate reductions from adopting the proposed method." @default.
- W3126996176 created "2021-02-15" @default.
- W3126996176 creator A5026649133 @default.
- W3126996176 creator A5037783894 @default.
- W3126996176 creator A5077473609 @default.
- W3126996176 date "2021-01-01" @default.
- W3126996176 modified "2023-10-13" @default.
- W3126996176 title "AG-SGD: Angle-Based Stochastic Gradient Descent" @default.
- W3126996176 cites W1988720110 @default.
- W3126996176 cites W2099940443 @default.
- W3126996176 cites W2240736559 @default.
- W3126996176 cites W2518029102 @default.
- W3126996176 cites W2744745684 @default.
- W3126996176 cites W2801995427 @default.
- W3126996176 cites W2899496144 @default.
- W3126996176 cites W2919936975 @default.
- W3126996176 cites W2942231644 @default.
- W3126996176 cites W2965857035 @default.
- W3126996176 cites W3002529007 @default.
- W3126996176 doi "https://doi.org/10.1109/access.2021.3055993" @default.
- W3126996176 hasPublicationYear "2021" @default.
- W3126996176 type Work @default.
- W3126996176 sameAs 3126996176 @default.
- W3126996176 citedByCount "3" @default.
- W3126996176 countsByYear W31269961762021 @default.
- W3126996176 countsByYear W31269961762022 @default.
- W3126996176 countsByYear W31269961762023 @default.
- W3126996176 crossrefType "journal-article" @default.
- W3126996176 hasAuthorship W3126996176A5026649133 @default.
- W3126996176 hasAuthorship W3126996176A5037783894 @default.
- W3126996176 hasAuthorship W3126996176A5077473609 @default.
- W3126996176 hasBestOaLocation W31269961761 @default.
- W3126996176 hasConcept C11413529 @default.
- W3126996176 hasConcept C115680565 @default.
- W3126996176 hasConcept C126255220 @default.
- W3126996176 hasConcept C127162648 @default.
- W3126996176 hasConcept C153258448 @default.
- W3126996176 hasConcept C154945302 @default.
- W3126996176 hasConcept C155032097 @default.
- W3126996176 hasConcept C162324750 @default.
- W3126996176 hasConcept C190502265 @default.
- W3126996176 hasConcept C206688291 @default.
- W3126996176 hasConcept C2777303404 @default.
- W3126996176 hasConcept C31258907 @default.
- W3126996176 hasConcept C33923547 @default.
- W3126996176 hasConcept C41008148 @default.
- W3126996176 hasConcept C45374587 @default.
- W3126996176 hasConcept C50522688 @default.
- W3126996176 hasConcept C50644808 @default.
- W3126996176 hasConcept C57869625 @default.
- W3126996176 hasConceptScore W3126996176C11413529 @default.
- W3126996176 hasConceptScore W3126996176C115680565 @default.
- W3126996176 hasConceptScore W3126996176C126255220 @default.
- W3126996176 hasConceptScore W3126996176C127162648 @default.
- W3126996176 hasConceptScore W3126996176C153258448 @default.
- W3126996176 hasConceptScore W3126996176C154945302 @default.
- W3126996176 hasConceptScore W3126996176C155032097 @default.
- W3126996176 hasConceptScore W3126996176C162324750 @default.
- W3126996176 hasConceptScore W3126996176C190502265 @default.
- W3126996176 hasConceptScore W3126996176C206688291 @default.
- W3126996176 hasConceptScore W3126996176C2777303404 @default.
- W3126996176 hasConceptScore W3126996176C31258907 @default.
- W3126996176 hasConceptScore W3126996176C33923547 @default.
- W3126996176 hasConceptScore W3126996176C41008148 @default.
- W3126996176 hasConceptScore W3126996176C45374587 @default.
- W3126996176 hasConceptScore W3126996176C50522688 @default.
- W3126996176 hasConceptScore W3126996176C50644808 @default.
- W3126996176 hasConceptScore W3126996176C57869625 @default.
- W3126996176 hasLocation W31269961761 @default.
- W3126996176 hasOpenAccess W3126996176 @default.
- W3126996176 hasPrimaryLocation W31269961761 @default.
- W3126996176 hasRelatedWork W1666485427 @default.
- W3126996176 hasRelatedWork W2766636532 @default.
- W3126996176 hasRelatedWork W2785612459 @default.
- W3126996176 hasRelatedWork W2979647660 @default.
- W3126996176 hasRelatedWork W3007279825 @default.
- W3126996176 hasRelatedWork W3098788150 @default.
- W3126996176 hasRelatedWork W3114199884 @default.
- W3126996176 hasRelatedWork W3126996176 @default.
- W3126996176 hasRelatedWork W3159389381 @default.
- W3126996176 hasRelatedWork W4295266589 @default.
- W3126996176 hasVolume "9" @default.
- W3126996176 isParatext "false" @default.
- W3126996176 isRetracted "false" @default.
- W3126996176 magId "3126996176" @default.
- W3126996176 workType "article" @default.