Matches in SemOpenAlex for { <https://semopenalex.org/work/W3113920149> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W3113920149 endingPage "10830" @default.
- W3113920149 startingPage "10820" @default.
- W3113920149 abstract "Recent empirical works show that large deep neural networks are often highly redundant and one can find much smaller subnetworks without a significant drop of accuracy. However, most existing methods of network pruning are empirical and heuristic, leaving it open whether good subnetworks provably exist, how to find them efficiently, and if network pruning can be provably better than direct training using gradient descent. We answer these problems positively by proposing a simple greedy selection approach for finding good subnetworks, which starts from an empty network and greedily adds important neurons from the large network. This differs from the existing methods based on backward elimination, which remove redundant neurons from the large network. Theoretically, applying the greedy selection strategy on sufficiently large {pre-trained} networks guarantees to find small subnetworks with lower loss than networks directly trained with gradient descent. Our results also apply to pruning randomly weighted networks. Practically, we improve prior arts of network pruning on learning compact neural architectures on ImageNet, including ResNet, MobilenetV2/V3, and ProxylessNet. Our theory and empirical results on MobileNet suggest that we should fine-tune the pruned subnetworks to leverage the information from the large model, instead of re-training from new random initialization as suggested in citet{liu2018rethinking}." @default.
- W3113920149 created "2021-01-05" @default.
- W3113920149 creator A5006626709 @default.
- W3113920149 creator A5011339261 @default.
- W3113920149 creator A5040717619 @default.
- W3113920149 creator A5057055806 @default.
- W3113920149 creator A5059919750 @default.
- W3113920149 creator A5061512999 @default.
- W3113920149 date "2020-11-21" @default.
- W3113920149 modified "2023-09-27" @default.
- W3113920149 title "Good Subnetworks Provably Exist: Pruning via Greedy Forward Selection" @default.
- W3113920149 hasPublicationYear "2020" @default.
- W3113920149 type Work @default.
- W3113920149 sameAs 3113920149 @default.
- W3113920149 citedByCount "18" @default.
- W3113920149 countsByYear W31139201492020 @default.
- W3113920149 countsByYear W31139201492021 @default.
- W3113920149 countsByYear W31139201492022 @default.
- W3113920149 crossrefType "proceedings-article" @default.
- W3113920149 hasAuthorship W3113920149A5006626709 @default.
- W3113920149 hasAuthorship W3113920149A5011339261 @default.
- W3113920149 hasAuthorship W3113920149A5040717619 @default.
- W3113920149 hasAuthorship W3113920149A5057055806 @default.
- W3113920149 hasAuthorship W3113920149A5059919750 @default.
- W3113920149 hasAuthorship W3113920149A5061512999 @default.
- W3113920149 hasConcept C108010975 @default.
- W3113920149 hasConcept C11413529 @default.
- W3113920149 hasConcept C114466953 @default.
- W3113920149 hasConcept C119857082 @default.
- W3113920149 hasConcept C153083717 @default.
- W3113920149 hasConcept C153258448 @default.
- W3113920149 hasConcept C154945302 @default.
- W3113920149 hasConcept C173801870 @default.
- W3113920149 hasConcept C199360897 @default.
- W3113920149 hasConcept C41008148 @default.
- W3113920149 hasConcept C50644808 @default.
- W3113920149 hasConcept C51823790 @default.
- W3113920149 hasConcept C6557445 @default.
- W3113920149 hasConcept C80444323 @default.
- W3113920149 hasConcept C81917197 @default.
- W3113920149 hasConcept C86803240 @default.
- W3113920149 hasConceptScore W3113920149C108010975 @default.
- W3113920149 hasConceptScore W3113920149C11413529 @default.
- W3113920149 hasConceptScore W3113920149C114466953 @default.
- W3113920149 hasConceptScore W3113920149C119857082 @default.
- W3113920149 hasConceptScore W3113920149C153083717 @default.
- W3113920149 hasConceptScore W3113920149C153258448 @default.
- W3113920149 hasConceptScore W3113920149C154945302 @default.
- W3113920149 hasConceptScore W3113920149C173801870 @default.
- W3113920149 hasConceptScore W3113920149C199360897 @default.
- W3113920149 hasConceptScore W3113920149C41008148 @default.
- W3113920149 hasConceptScore W3113920149C50644808 @default.
- W3113920149 hasConceptScore W3113920149C51823790 @default.
- W3113920149 hasConceptScore W3113920149C6557445 @default.
- W3113920149 hasConceptScore W3113920149C80444323 @default.
- W3113920149 hasConceptScore W3113920149C81917197 @default.
- W3113920149 hasConceptScore W3113920149C86803240 @default.
- W3113920149 hasLocation W31139201491 @default.
- W3113920149 hasOpenAccess W3113920149 @default.
- W3113920149 hasPrimaryLocation W31139201491 @default.
- W3113920149 hasRelatedWork W1686810756 @default.
- W3113920149 hasRelatedWork W1821462560 @default.
- W3113920149 hasRelatedWork W2114766824 @default.
- W3113920149 hasRelatedWork W2194775991 @default.
- W3113920149 hasRelatedWork W2808168148 @default.
- W3113920149 hasRelatedWork W2886851211 @default.
- W3113920149 hasRelatedWork W2896409484 @default.
- W3113920149 hasRelatedWork W2928560789 @default.
- W3113920149 hasRelatedWork W2948635472 @default.
- W3113920149 hasRelatedWork W2962851801 @default.
- W3113920149 hasRelatedWork W2962965870 @default.
- W3113920149 hasRelatedWork W2963363373 @default.
- W3113920149 hasRelatedWork W2963674932 @default.
- W3113920149 hasRelatedWork W2963813662 @default.
- W3113920149 hasRelatedWork W2964233199 @default.
- W3113920149 hasRelatedWork W2964299589 @default.
- W3113920149 hasRelatedWork W3009043942 @default.
- W3113920149 hasRelatedWork W3034513523 @default.
- W3113920149 hasRelatedWork W3035615218 @default.
- W3113920149 hasRelatedWork W3118608800 @default.
- W3113920149 isParatext "false" @default.
- W3113920149 isRetracted "false" @default.
- W3113920149 magId "3113920149" @default.
- W3113920149 workType "article" @default.