Matches in SemOpenAlex for { <https://semopenalex.org/work/W4321610099> ?p ?o ?g. }
Showing items 1 to 95 of
95
with 100 items per page.
- W4321610099 endingPage "e0276427" @default.
- W4321610099 startingPage "e0276427" @default.
- W4321610099 abstract "To break the three lockings during backpropagation (BP) process for neural network training, multiple decoupled learning methods have been investigated recently. These methods either lead to significant drop in accuracy performance or suffer from dramatic increase in memory usage. In this paper, a new form of decoupled learning, named decoupled neural network training scheme with re-computation and weight prediction (DTRP) is proposed. In DTRP, a re-computation scheme is adopted to solve the memory explosion problem, and a weight prediction scheme is proposed to deal with the weight delay caused by re-computation. Additionally, a batch compensation scheme is developed, allowing the proposed DTRP to run faster. Theoretical analysis shows that DTRP is guaranteed to converge to crical points under certain conditions. Experiments are conducted by training various convolutional neural networks on several classification datasets, showing comparable or better results than the state-of-the-art methods and BP. These experiments also reveal that adopting the proposed method, the memory explosion problem is effectively solved, and a significant acceleration is achieved." @default.
- W4321610099 created "2023-02-24" @default.
- W4321610099 creator A5022066559 @default.
- W4321610099 creator A5049506273 @default.
- W4321610099 creator A5059175011 @default.
- W4321610099 creator A5061256037 @default.
- W4321610099 creator A5067922510 @default.
- W4321610099 creator A5073783353 @default.
- W4321610099 date "2023-02-23" @default.
- W4321610099 modified "2023-10-14" @default.
- W4321610099 title "Decoupled neural network training with re-computation and weight prediction" @default.
- W4321610099 cites W1498436455 @default.
- W4321610099 cites W1677182931 @default.
- W4321610099 cites W1994616650 @default.
- W4321610099 cites W2108598243 @default.
- W4321610099 cites W2194775991 @default.
- W4321610099 cites W2583832915 @default.
- W4321610099 cites W2612806369 @default.
- W4321610099 cites W2741295441 @default.
- W4321610099 cites W2943370233 @default.
- W4321610099 cites W2947411064 @default.
- W4321610099 cites W2949933418 @default.
- W4321610099 cites W2963433607 @default.
- W4321610099 cites W2963446712 @default.
- W4321610099 cites W2964319207 @default.
- W4321610099 cites W3119315021 @default.
- W4321610099 cites W3143471131 @default.
- W4321610099 cites W3176923149 @default.
- W4321610099 doi "https://doi.org/10.1371/journal.pone.0276427" @default.
- W4321610099 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/36821537" @default.
- W4321610099 hasPublicationYear "2023" @default.
- W4321610099 type Work @default.
- W4321610099 citedByCount "0" @default.
- W4321610099 crossrefType "journal-article" @default.
- W4321610099 hasAuthorship W4321610099A5022066559 @default.
- W4321610099 hasAuthorship W4321610099A5049506273 @default.
- W4321610099 hasAuthorship W4321610099A5059175011 @default.
- W4321610099 hasAuthorship W4321610099A5061256037 @default.
- W4321610099 hasAuthorship W4321610099A5067922510 @default.
- W4321610099 hasAuthorship W4321610099A5073783353 @default.
- W4321610099 hasBestOaLocation W43216100991 @default.
- W4321610099 hasConcept C111919701 @default.
- W4321610099 hasConcept C11413529 @default.
- W4321610099 hasConcept C117896860 @default.
- W4321610099 hasConcept C119857082 @default.
- W4321610099 hasConcept C121332964 @default.
- W4321610099 hasConcept C134306372 @default.
- W4321610099 hasConcept C154945302 @default.
- W4321610099 hasConcept C155032097 @default.
- W4321610099 hasConcept C33923547 @default.
- W4321610099 hasConcept C41008148 @default.
- W4321610099 hasConcept C45374587 @default.
- W4321610099 hasConcept C50644808 @default.
- W4321610099 hasConcept C74650414 @default.
- W4321610099 hasConcept C77618280 @default.
- W4321610099 hasConcept C81363708 @default.
- W4321610099 hasConcept C98045186 @default.
- W4321610099 hasConceptScore W4321610099C111919701 @default.
- W4321610099 hasConceptScore W4321610099C11413529 @default.
- W4321610099 hasConceptScore W4321610099C117896860 @default.
- W4321610099 hasConceptScore W4321610099C119857082 @default.
- W4321610099 hasConceptScore W4321610099C121332964 @default.
- W4321610099 hasConceptScore W4321610099C134306372 @default.
- W4321610099 hasConceptScore W4321610099C154945302 @default.
- W4321610099 hasConceptScore W4321610099C155032097 @default.
- W4321610099 hasConceptScore W4321610099C33923547 @default.
- W4321610099 hasConceptScore W4321610099C41008148 @default.
- W4321610099 hasConceptScore W4321610099C45374587 @default.
- W4321610099 hasConceptScore W4321610099C50644808 @default.
- W4321610099 hasConceptScore W4321610099C74650414 @default.
- W4321610099 hasConceptScore W4321610099C77618280 @default.
- W4321610099 hasConceptScore W4321610099C81363708 @default.
- W4321610099 hasConceptScore W4321610099C98045186 @default.
- W4321610099 hasIssue "2" @default.
- W4321610099 hasLocation W43216100991 @default.
- W4321610099 hasLocation W43216100992 @default.
- W4321610099 hasLocation W43216100993 @default.
- W4321610099 hasOpenAccess W4321610099 @default.
- W4321610099 hasPrimaryLocation W43216100991 @default.
- W4321610099 hasRelatedWork W1975341817 @default.
- W4321610099 hasRelatedWork W2040019480 @default.
- W4321610099 hasRelatedWork W2386707516 @default.
- W4321610099 hasRelatedWork W2391384657 @default.
- W4321610099 hasRelatedWork W2894173309 @default.
- W4321610099 hasRelatedWork W3021430260 @default.
- W4321610099 hasRelatedWork W3027997911 @default.
- W4321610099 hasRelatedWork W3156786002 @default.
- W4321610099 hasRelatedWork W4287776258 @default.
- W4321610099 hasRelatedWork W4304136584 @default.
- W4321610099 hasVolume "18" @default.
- W4321610099 isParatext "false" @default.
- W4321610099 isRetracted "false" @default.
- W4321610099 workType "article" @default.