Matches in SemOpenAlex for { <https://semopenalex.org/work/W4225567919> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4225567919 abstract "Neural network pruning has remarkable performance for reducing the complexity of deep network models. Recent network pruning methods usually focused on removing unimportant or redundant filters in the network. In this paper, by exploring the similarities between feature maps, we propose a novel filter pruning method, Central Filter (CF), which suggests that a filter is approximately equal to a set of other filters after appropriate adjustments. Our method is based on the discovery that the average similarity between feature maps changes very little, regardless of the number of input images. Based on this finding, we establish similarity graphs on feature maps and calculate the closeness centrality of each node to select the Central Filter. Moreover, we design a method to directly adjust weights in the next layer corresponding to the Central Filter, effectively minimizing the error caused by pruning. Through experiments on various benchmark networks and datasets, CF yields state-of-the-art performance. For example, with ResNet-56, CF reduces approximately 39.7% of FLOPs by removing 47.1% of the parameters, with even 0.33% accuracy improvement on CIFAR-10. With GoogLeNet, CF reduces approximately 63.2% of FLOPs by removing 55.6% of the parameters, with only a small loss of 0.35% in top-1 accuracy on CIFAR-10. With ResNet-50, CF reduces approximately 47.9% of FLOPs by removing 36.9% of the parameters, with only a small loss of 1.07% in top-1 accuracy on ImageNet. The codes can be available at https://github.com/8ubpshLR23/Central-Filter." @default.
- W4225567919 created "2022-05-05" @default.
- W4225567919 creator A5035048973 @default.
- W4225567919 creator A5057055806 @default.
- W4225567919 creator A5072152983 @default.
- W4225567919 creator A5085506535 @default.
- W4225567919 creator A5087454088 @default.
- W4225567919 date "2021-12-10" @default.
- W4225567919 modified "2023-09-27" @default.
- W4225567919 title "Network Compression via Central Filter" @default.
- W4225567919 doi "https://doi.org/10.48550/arxiv.2112.05493" @default.
- W4225567919 hasPublicationYear "2021" @default.
- W4225567919 type Work @default.
- W4225567919 citedByCount "0" @default.
- W4225567919 crossrefType "posted-content" @default.
- W4225567919 hasAuthorship W4225567919A5035048973 @default.
- W4225567919 hasAuthorship W4225567919A5057055806 @default.
- W4225567919 hasAuthorship W4225567919A5072152983 @default.
- W4225567919 hasAuthorship W4225567919A5085506535 @default.
- W4225567919 hasAuthorship W4225567919A5087454088 @default.
- W4225567919 hasBestOaLocation W42255679191 @default.
- W4225567919 hasConcept C103278499 @default.
- W4225567919 hasConcept C106131492 @default.
- W4225567919 hasConcept C108010975 @default.
- W4225567919 hasConcept C11413529 @default.
- W4225567919 hasConcept C115961682 @default.
- W4225567919 hasConcept C13280743 @default.
- W4225567919 hasConcept C138885662 @default.
- W4225567919 hasConcept C153180895 @default.
- W4225567919 hasConcept C154945302 @default.
- W4225567919 hasConcept C173608175 @default.
- W4225567919 hasConcept C185798385 @default.
- W4225567919 hasConcept C205649164 @default.
- W4225567919 hasConcept C2776401178 @default.
- W4225567919 hasConcept C31972630 @default.
- W4225567919 hasConcept C3826847 @default.
- W4225567919 hasConcept C41008148 @default.
- W4225567919 hasConcept C41895202 @default.
- W4225567919 hasConcept C50644808 @default.
- W4225567919 hasConcept C6557445 @default.
- W4225567919 hasConcept C86803240 @default.
- W4225567919 hasConceptScore W4225567919C103278499 @default.
- W4225567919 hasConceptScore W4225567919C106131492 @default.
- W4225567919 hasConceptScore W4225567919C108010975 @default.
- W4225567919 hasConceptScore W4225567919C11413529 @default.
- W4225567919 hasConceptScore W4225567919C115961682 @default.
- W4225567919 hasConceptScore W4225567919C13280743 @default.
- W4225567919 hasConceptScore W4225567919C138885662 @default.
- W4225567919 hasConceptScore W4225567919C153180895 @default.
- W4225567919 hasConceptScore W4225567919C154945302 @default.
- W4225567919 hasConceptScore W4225567919C173608175 @default.
- W4225567919 hasConceptScore W4225567919C185798385 @default.
- W4225567919 hasConceptScore W4225567919C205649164 @default.
- W4225567919 hasConceptScore W4225567919C2776401178 @default.
- W4225567919 hasConceptScore W4225567919C31972630 @default.
- W4225567919 hasConceptScore W4225567919C3826847 @default.
- W4225567919 hasConceptScore W4225567919C41008148 @default.
- W4225567919 hasConceptScore W4225567919C41895202 @default.
- W4225567919 hasConceptScore W4225567919C50644808 @default.
- W4225567919 hasConceptScore W4225567919C6557445 @default.
- W4225567919 hasConceptScore W4225567919C86803240 @default.
- W4225567919 hasLocation W42255679191 @default.
- W4225567919 hasOpenAccess W4225567919 @default.
- W4225567919 hasPrimaryLocation W42255679191 @default.
- W4225567919 hasRelatedWork W2808568129 @default.
- W4225567919 hasRelatedWork W2927393429 @default.
- W4225567919 hasRelatedWork W2956095049 @default.
- W4225567919 hasRelatedWork W2970216048 @default.
- W4225567919 hasRelatedWork W3008861477 @default.
- W4225567919 hasRelatedWork W3034513523 @default.
- W4225567919 hasRelatedWork W3094522247 @default.
- W4225567919 hasRelatedWork W3119066197 @default.
- W4225567919 hasRelatedWork W4200341551 @default.
- W4225567919 hasRelatedWork W4207000934 @default.
- W4225567919 isParatext "false" @default.
- W4225567919 isRetracted "false" @default.
- W4225567919 workType "article" @default.