FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313015248> ?p ?o ?g. }

• W4313015248 endingPage "2199" @default.
• W4313015248 startingPage "2187" @default.
• W4313015248 abstract "Processing-in-memory (PIM) enables energy-efficient deployment of convolutional neural networks (CNNs) from edge to cloud. Resistive random-access memory (ReRAM) is one of the most commonly used technologies for PIM architectures. One of the primary limitations of ReRAM-based PIM in neural network training arises from the limited write endurance due to the frequent weight updates. To make ReRAM-based architectures viable for CNN training, the write endurance issue needs to be addressed. This work aims to reduce the number of weight reprogrammings without compromising the final model accuracy. We propose the ESSENCE framework with an endurance-aware structured stochastic gradient pruning method, which dynamically adjusts the probability of gradient update based on the current update counts. Experimental results with multiple CNNs and datasets demonstrate that the proposed method can extend ReRAM’s life time for training. For instance, with the ResNet20 network and CIFAR-10 dataset, ESSENCE can save the mean update counts of up to <inline-formula xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink> <tex-math notation=LaTeX>$10.29times $ </tex-math></inline-formula> compared to the stochastic gradient descent method and effectively reduce the maximum update counts compared with the No Endurance method. Furthermore, an aggressive tuning method based on ESSENCE can boost the mean update count savings by up to <inline-formula xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink> <tex-math notation=LaTeX>$14.41times $ </tex-math></inline-formula> ." @default.
• W4313015248 created "2023-01-05" @default.
• W4313015248 creator A5033880864 @default.
• W4313015248 creator A5055445718 @default.
• W4313015248 creator A5067763720 @default.
• W4313015248 creator A5075474462 @default.
• W4313015248 creator A5076154259 @default.
• W4313015248 creator A5080236384 @default.
• W4313015248 date "2023-07-01" @default.
• W4313015248 modified "2023-10-18" @default.
• W4313015248 title "ESSENCE: Exploiting Structured Stochastic Gradient Pruning for Endurance-Aware ReRAM-Based In-Memory Training Systems" @default.
• W4313015248 cites W114517082 @default.
• W4313015248 cites W1517256516 @default.
• W4313015248 cites W1976075132 @default.
• W4313015248 cites W1980287119 @default.
• W4313015248 cites W2036899386 @default.
• W4313015248 cites W2194775991 @default.
• W4313015248 cites W2508602506 @default.
• W4313015248 cites W2518281301 @default.
• W4313015248 cites W2545405476 @default.
• W4313015248 cites W2555487162 @default.
• W4313015248 cites W2556936974 @default.
• W4313015248 cites W2613989746 @default.
• W4313015248 cites W2799229073 @default.
• W4313015248 cites W2809617027 @default.
• W4313015248 cites W2846439410 @default.
• W4313015248 cites W2946047477 @default.
• W4313015248 cites W2946591837 @default.
• W4313015248 cites W3000505821 @default.
• W4313015248 cites W3005999147 @default.
• W4313015248 cites W3107472389 @default.
• W4313015248 cites W3111375540 @default.
• W4313015248 cites W3185576717 @default.
• W4313015248 cites W3201198776 @default.
• W4313015248 cites W3201394727 @default.
• W4313015248 cites W4225679468 @default.
• W4313015248 cites W4226070519 @default.
• W4313015248 cites W4232753305 @default.
• W4313015248 cites W4245731639 @default.
• W4313015248 cites W4281289797 @default.
• W4313015248 cites W4282829327 @default.
• W4313015248 doi "https://doi.org/10.1109/tcad.2022.3216546" @default.
• W4313015248 hasPublicationYear "2023" @default.
• W4313015248 type Work @default.
• W4313015248 citedByCount "1" @default.
• W4313015248 countsByYear W43130152482022 @default.
• W4313015248 crossrefType "journal-article" @default.
• W4313015248 hasAuthorship W4313015248A5033880864 @default.
• W4313015248 hasAuthorship W4313015248A5055445718 @default.
• W4313015248 hasAuthorship W4313015248A5067763720 @default.
• W4313015248 hasAuthorship W4313015248A5075474462 @default.
• W4313015248 hasAuthorship W4313015248A5076154259 @default.
• W4313015248 hasAuthorship W4313015248A5080236384 @default.
• W4313015248 hasConcept C108010975 @default.
• W4313015248 hasConcept C11413529 @default.
• W4313015248 hasConcept C119599485 @default.
• W4313015248 hasConcept C119857082 @default.
• W4313015248 hasConcept C127413603 @default.
• W4313015248 hasConcept C154945302 @default.
• W4313015248 hasConcept C165801399 @default.
• W4313015248 hasConcept C182019814 @default.
• W4313015248 hasConcept C206688291 @default.
• W4313015248 hasConcept C33923547 @default.
• W4313015248 hasConcept C41008148 @default.
• W4313015248 hasConcept C45357846 @default.
• W4313015248 hasConcept C50644808 @default.
• W4313015248 hasConcept C6557445 @default.
• W4313015248 hasConcept C81363708 @default.
• W4313015248 hasConcept C86803240 @default.
• W4313015248 hasConcept C94375191 @default.
• W4313015248 hasConceptScore W4313015248C108010975 @default.
• W4313015248 hasConceptScore W4313015248C11413529 @default.
• W4313015248 hasConceptScore W4313015248C119599485 @default.
• W4313015248 hasConceptScore W4313015248C119857082 @default.
• W4313015248 hasConceptScore W4313015248C127413603 @default.
• W4313015248 hasConceptScore W4313015248C154945302 @default.
• W4313015248 hasConceptScore W4313015248C165801399 @default.
• W4313015248 hasConceptScore W4313015248C182019814 @default.
• W4313015248 hasConceptScore W4313015248C206688291 @default.
• W4313015248 hasConceptScore W4313015248C33923547 @default.
• W4313015248 hasConceptScore W4313015248C41008148 @default.
• W4313015248 hasConceptScore W4313015248C45357846 @default.
• W4313015248 hasConceptScore W4313015248C50644808 @default.
• W4313015248 hasConceptScore W4313015248C6557445 @default.
• W4313015248 hasConceptScore W4313015248C81363708 @default.
• W4313015248 hasConceptScore W4313015248C86803240 @default.
• W4313015248 hasConceptScore W4313015248C94375191 @default.
• W4313015248 hasFunder F4320306076 @default.
• W4313015248 hasIssue "7" @default.
• W4313015248 hasLocation W43130152481 @default.
• W4313015248 hasOpenAccess W4313015248 @default.
• W4313015248 hasPrimaryLocation W43130152481 @default.
• W4313015248 hasRelatedWork W2748454020 @default.
• W4313015248 hasRelatedWork W2787191226 @default.
• W4313015248 hasRelatedWork W2792987183 @default.
• W4313015248 hasRelatedWork W2961085424 @default.
• W4313015248 hasRelatedWork W2989932438 @default.
• W4313015248 hasRelatedWork W3021430260 @default.

• next