Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386764067> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W4386764067 abstract "Convolutional neural networks (CNNs) are popular in high-performing solutions to many real-world problems, such as audio classification. CNNs have many parameters and filters, with some having a larger impact on the performance than others. This means that networks may contain many unnecessary filters, increasing a CNN’s computation and memory requirements while providing limited performance benefits. To make CNNs more efficient, we propose a pruning framework that eliminates filters with the highest commonality. We measure this commonality using the graph-theoretic concept of centrality. We hypothesise that a filter with a high centrality should be eliminated as it represents commonality and can be replaced by other filters without affecting the performance of a network much. An experimental evaluation of the proposed framework is performed on acoustic scene classification and audio tagging. On the DCASE 2021 Task 1A baseline network, our proposed method reduces computations per inference by 71% with 50% fewer parameters with less than a two percentage point drop in accuracy compared to the original network. For large-scale CNNs such as PANNs designed for audio tagging, our method reduces computations per inference by 24% with 41% fewer parameters at a slight improvement in performance." @default.
- W4386764067 created "2023-09-16" @default.
- W4386764067 creator A5007681214 @default.
- W4386764067 creator A5011366980 @default.
- W4386764067 creator A5066967599 @default.
- W4386764067 date "2023-10-22" @default.
- W4386764067 modified "2023-09-27" @default.
- W4386764067 title "Compressing Audio CNNS with Graph Centrality Based Filter Pruning" @default.
- W4386764067 cites W2155307989 @default.
- W4386764067 cites W2593116425 @default.
- W4386764067 cites W2924515500 @default.
- W4386764067 cites W2928560789 @default.
- W4386764067 cites W2931364255 @default.
- W4386764067 cites W2962851801 @default.
- W4386764067 cites W2972592847 @default.
- W4386764067 cites W2975625678 @default.
- W4386764067 cites W3034513523 @default.
- W4386764067 cites W3094550259 @default.
- W4386764067 cites W3100097016 @default.
- W4386764067 cites W3184606595 @default.
- W4386764067 cites W4225137843 @default.
- W4386764067 cites W4226126941 @default.
- W4386764067 doi "https://doi.org/10.1109/waspaa58266.2023.10248103" @default.
- W4386764067 hasPublicationYear "2023" @default.
- W4386764067 type Work @default.
- W4386764067 citedByCount "0" @default.
- W4386764067 crossrefType "proceedings-article" @default.
- W4386764067 hasAuthorship W4386764067A5007681214 @default.
- W4386764067 hasAuthorship W4386764067A5011366980 @default.
- W4386764067 hasAuthorship W4386764067A5066967599 @default.
- W4386764067 hasConcept C106131492 @default.
- W4386764067 hasConcept C108010975 @default.
- W4386764067 hasConcept C11413529 @default.
- W4386764067 hasConcept C114614502 @default.
- W4386764067 hasConcept C119857082 @default.
- W4386764067 hasConcept C132525143 @default.
- W4386764067 hasConcept C153180895 @default.
- W4386764067 hasConcept C154945302 @default.
- W4386764067 hasConcept C2776214188 @default.
- W4386764067 hasConcept C31972630 @default.
- W4386764067 hasConcept C33923547 @default.
- W4386764067 hasConcept C41008148 @default.
- W4386764067 hasConcept C45374587 @default.
- W4386764067 hasConcept C53811970 @default.
- W4386764067 hasConcept C6557445 @default.
- W4386764067 hasConcept C80444323 @default.
- W4386764067 hasConcept C81363708 @default.
- W4386764067 hasConcept C86803240 @default.
- W4386764067 hasConceptScore W4386764067C106131492 @default.
- W4386764067 hasConceptScore W4386764067C108010975 @default.
- W4386764067 hasConceptScore W4386764067C11413529 @default.
- W4386764067 hasConceptScore W4386764067C114614502 @default.
- W4386764067 hasConceptScore W4386764067C119857082 @default.
- W4386764067 hasConceptScore W4386764067C132525143 @default.
- W4386764067 hasConceptScore W4386764067C153180895 @default.
- W4386764067 hasConceptScore W4386764067C154945302 @default.
- W4386764067 hasConceptScore W4386764067C2776214188 @default.
- W4386764067 hasConceptScore W4386764067C31972630 @default.
- W4386764067 hasConceptScore W4386764067C33923547 @default.
- W4386764067 hasConceptScore W4386764067C41008148 @default.
- W4386764067 hasConceptScore W4386764067C45374587 @default.
- W4386764067 hasConceptScore W4386764067C53811970 @default.
- W4386764067 hasConceptScore W4386764067C6557445 @default.
- W4386764067 hasConceptScore W4386764067C80444323 @default.
- W4386764067 hasConceptScore W4386764067C81363708 @default.
- W4386764067 hasConceptScore W4386764067C86803240 @default.
- W4386764067 hasFunder F4320334627 @default.
- W4386764067 hasLocation W43867640671 @default.
- W4386764067 hasOpenAccess W4386764067 @default.
- W4386764067 hasPrimaryLocation W43867640671 @default.
- W4386764067 hasRelatedWork W2767651786 @default.
- W4386764067 hasRelatedWork W2912288872 @default.
- W4386764067 hasRelatedWork W2941817504 @default.
- W4386764067 hasRelatedWork W2944261608 @default.
- W4386764067 hasRelatedWork W2997006882 @default.
- W4386764067 hasRelatedWork W3021430260 @default.
- W4386764067 hasRelatedWork W3027997911 @default.
- W4386764067 hasRelatedWork W4231283761 @default.
- W4386764067 hasRelatedWork W4287776258 @default.
- W4386764067 hasRelatedWork W4312358390 @default.
- W4386764067 isParatext "false" @default.
- W4386764067 isRetracted "false" @default.
- W4386764067 workType "article" @default.