Matches in SemOpenAlex for { <https://semopenalex.org/work/W3104920644> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W3104920644 endingPage "14080" @default.
- W3104920644 startingPage "14068" @default.
- W3104920644 abstract "Scaling up the convolutional neural network (CNN) size (e.g., width, depth, etc.) is known to effectively improve model accuracy. However, the large model size impedes training on resource-constrained edge devices. For instance, federated learning (FL) may place undue burden on the compute capability of edge nodes, even though there is a strong practical need for FL due to its privacy and confidentiality properties. To address the resource-constrained reality of edge devices, we reformulate FL as a group knowledge transfer training algorithm, called FedGKT. FedGKT designs a variant of the alternating minimization approach to train small CNNs on edge nodes and periodically transfer their knowledge by knowledge distillation to a large server-side CNN. FedGKT consolidates several advantages into a single framework: reduced demand for edge computation, lower communication bandwidth for large CNNs, and asynchronous training, all while maintaining model accuracy comparable to FedAvg. We train CNNs designed based on ResNet-56 and ResNet-110 using three distinct datasets (CIFAR-10, CIFAR-100, and CINIC-10) and their non-I.I.D. variants. Our results show that FedGKT can obtain comparable or even slightly higher accuracy than FedAvg. More importantly, FedGKT makes edge training affordable. Compared to the edge training using FedAvg, FedGKT demands 9 to 17 times less computational power (FLOPs) on edge devices and requires 54 to 105 times fewer parameters in the edge CNN. Our source code is released at FedML (this https URL)." @default.
- W3104920644 created "2020-11-23" @default.
- W3104920644 creator A5003518978 @default.
- W3104920644 creator A5018033573 @default.
- W3104920644 creator A5034183260 @default.
- W3104920644 date "2020-01-01" @default.
- W3104920644 modified "2023-10-18" @default.
- W3104920644 title "Group Knowledge Transfer: Federated Learning of Large CNNs at the Edge" @default.
- W3104920644 hasPublicationYear "2020" @default.
- W3104920644 type Work @default.
- W3104920644 sameAs 3104920644 @default.
- W3104920644 citedByCount "27" @default.
- W3104920644 countsByYear W31049206442019 @default.
- W3104920644 countsByYear W31049206442020 @default.
- W3104920644 countsByYear W31049206442021 @default.
- W3104920644 countsByYear W31049206442022 @default.
- W3104920644 crossrefType "proceedings-article" @default.
- W3104920644 hasAuthorship W3104920644A5003518978 @default.
- W3104920644 hasAuthorship W3104920644A5018033573 @default.
- W3104920644 hasAuthorship W3104920644A5034183260 @default.
- W3104920644 hasConcept C108583219 @default.
- W3104920644 hasConcept C111919701 @default.
- W3104920644 hasConcept C113775141 @default.
- W3104920644 hasConcept C11413529 @default.
- W3104920644 hasConcept C119857082 @default.
- W3104920644 hasConcept C138236772 @default.
- W3104920644 hasConcept C150899416 @default.
- W3104920644 hasConcept C154945302 @default.
- W3104920644 hasConcept C162307627 @default.
- W3104920644 hasConcept C2778456923 @default.
- W3104920644 hasConcept C41008148 @default.
- W3104920644 hasConcept C45374587 @default.
- W3104920644 hasConcept C79974875 @default.
- W3104920644 hasConcept C81363708 @default.
- W3104920644 hasConceptScore W3104920644C108583219 @default.
- W3104920644 hasConceptScore W3104920644C111919701 @default.
- W3104920644 hasConceptScore W3104920644C113775141 @default.
- W3104920644 hasConceptScore W3104920644C11413529 @default.
- W3104920644 hasConceptScore W3104920644C119857082 @default.
- W3104920644 hasConceptScore W3104920644C138236772 @default.
- W3104920644 hasConceptScore W3104920644C150899416 @default.
- W3104920644 hasConceptScore W3104920644C154945302 @default.
- W3104920644 hasConceptScore W3104920644C162307627 @default.
- W3104920644 hasConceptScore W3104920644C2778456923 @default.
- W3104920644 hasConceptScore W3104920644C41008148 @default.
- W3104920644 hasConceptScore W3104920644C45374587 @default.
- W3104920644 hasConceptScore W3104920644C79974875 @default.
- W3104920644 hasConceptScore W3104920644C81363708 @default.
- W3104920644 hasLocation W31049206441 @default.
- W3104920644 hasOpenAccess W3104920644 @default.
- W3104920644 hasPrimaryLocation W31049206441 @default.
- W3104920644 hasRelatedWork W1821462560 @default.
- W3104920644 hasRelatedWork W2194775991 @default.
- W3104920644 hasRelatedWork W2535838896 @default.
- W3104920644 hasRelatedWork W2541884796 @default.
- W3104920644 hasRelatedWork W2807006176 @default.
- W3104920644 hasRelatedWork W2902113386 @default.
- W3104920644 hasRelatedWork W2912213068 @default.
- W3104920644 hasRelatedWork W2962688627 @default.
- W3104920644 hasRelatedWork W2963300197 @default.
- W3104920644 hasRelatedWork W2976335444 @default.
- W3104920644 hasRelatedWork W2980216952 @default.
- W3104920644 hasRelatedWork W2995022099 @default.
- W3104920644 hasRelatedWork W3006017224 @default.
- W3104920644 hasRelatedWork W3008187686 @default.
- W3104920644 hasRelatedWork W3034666238 @default.
- W3104920644 hasRelatedWork W3038022836 @default.
- W3104920644 hasRelatedWork W3044211235 @default.
- W3104920644 hasRelatedWork W3100393648 @default.
- W3104920644 hasRelatedWork W3103802018 @default.
- W3104920644 hasRelatedWork W3118608800 @default.
- W3104920644 hasVolume "33" @default.
- W3104920644 isParatext "false" @default.
- W3104920644 isRetracted "false" @default.
- W3104920644 magId "3104920644" @default.
- W3104920644 workType "article" @default.