Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100809954> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W3100809954 abstract "The sizes of deep neural networks (DNNs) are rapidly outgrowing the capacity of hardware to store and train them. Research over the past few decades has explored the prospect of sparsifying DNNs before, during, and after training by pruning edges from the underlying topology. The resulting neural network is known as a sparse neural network. More recent work has demonstrated the remarkable result that certain sparse DNNs can train to the same precision as dense DNNs at lower runtime and storage cost. An intriguing class of these sparse DNNs is the X-Nets, which are initialized and trained upon a sparse topology with neither reference to a parent dense DNN nor subsequent pruning. We present an algorithm that deterministically generates RadiX-Nets: sparse DNN topologies that, as a whole, are much more diverse than X-Net topologies, while preserving X-Nets' desired characteristics. We further present a functional-analytic conjecture based on the longstanding observation that sparse neural network topologies can attain the same expressive power as dense counterparts" @default.
- W3100809954 created "2020-11-23" @default.
- W3100809954 creator A5003145273 @default.
- W3100809954 creator A5072108599 @default.
- W3100809954 date "2019-05-01" @default.
- W3100809954 modified "2023-09-28" @default.
- W3100809954 title "RadiX-Net: Structured Sparse Matrices for Deep Neural Networks" @default.
- W3100809954 cites W2073389244 @default.
- W3100809954 cites W2097117768 @default.
- W3100809954 cites W2103496339 @default.
- W3100809954 cites W26556108 @default.
- W3100809954 cites W2809429607 @default.
- W3100809954 cites W4233079698 @default.
- W3100809954 cites W4240805545 @default.
- W3100809954 cites W2859678021 @default.
- W3100809954 doi "https://doi.org/10.1109/ipdpsw.2019.00051" @default.
- W3100809954 hasPublicationYear "2019" @default.
- W3100809954 type Work @default.
- W3100809954 sameAs 3100809954 @default.
- W3100809954 citedByCount "21" @default.
- W3100809954 countsByYear W31008099542020 @default.
- W3100809954 countsByYear W31008099542021 @default.
- W3100809954 countsByYear W31008099542022 @default.
- W3100809954 countsByYear W31008099542023 @default.
- W3100809954 crossrefType "proceedings-article" @default.
- W3100809954 hasAuthorship W3100809954A5003145273 @default.
- W3100809954 hasAuthorship W3100809954A5072108599 @default.
- W3100809954 hasBestOaLocation W31008099542 @default.
- W3100809954 hasConcept C108010975 @default.
- W3100809954 hasConcept C11413529 @default.
- W3100809954 hasConcept C114614502 @default.
- W3100809954 hasConcept C121332964 @default.
- W3100809954 hasConcept C14166107 @default.
- W3100809954 hasConcept C154945302 @default.
- W3100809954 hasConcept C163716315 @default.
- W3100809954 hasConcept C184720557 @default.
- W3100809954 hasConcept C195818886 @default.
- W3100809954 hasConcept C199845137 @default.
- W3100809954 hasConcept C2524010 @default.
- W3100809954 hasConcept C2777212361 @default.
- W3100809954 hasConcept C2984842247 @default.
- W3100809954 hasConcept C31258907 @default.
- W3100809954 hasConcept C33923547 @default.
- W3100809954 hasConcept C41008148 @default.
- W3100809954 hasConcept C50644808 @default.
- W3100809954 hasConcept C56372850 @default.
- W3100809954 hasConcept C62520636 @default.
- W3100809954 hasConcept C6557445 @default.
- W3100809954 hasConcept C80444323 @default.
- W3100809954 hasConcept C86803240 @default.
- W3100809954 hasConceptScore W3100809954C108010975 @default.
- W3100809954 hasConceptScore W3100809954C11413529 @default.
- W3100809954 hasConceptScore W3100809954C114614502 @default.
- W3100809954 hasConceptScore W3100809954C121332964 @default.
- W3100809954 hasConceptScore W3100809954C14166107 @default.
- W3100809954 hasConceptScore W3100809954C154945302 @default.
- W3100809954 hasConceptScore W3100809954C163716315 @default.
- W3100809954 hasConceptScore W3100809954C184720557 @default.
- W3100809954 hasConceptScore W3100809954C195818886 @default.
- W3100809954 hasConceptScore W3100809954C199845137 @default.
- W3100809954 hasConceptScore W3100809954C2524010 @default.
- W3100809954 hasConceptScore W3100809954C2777212361 @default.
- W3100809954 hasConceptScore W3100809954C2984842247 @default.
- W3100809954 hasConceptScore W3100809954C31258907 @default.
- W3100809954 hasConceptScore W3100809954C33923547 @default.
- W3100809954 hasConceptScore W3100809954C41008148 @default.
- W3100809954 hasConceptScore W3100809954C50644808 @default.
- W3100809954 hasConceptScore W3100809954C56372850 @default.
- W3100809954 hasConceptScore W3100809954C62520636 @default.
- W3100809954 hasConceptScore W3100809954C6557445 @default.
- W3100809954 hasConceptScore W3100809954C80444323 @default.
- W3100809954 hasConceptScore W3100809954C86803240 @default.
- W3100809954 hasLocation W31008099541 @default.
- W3100809954 hasLocation W31008099542 @default.
- W3100809954 hasOpenAccess W3100809954 @default.
- W3100809954 hasPrimaryLocation W31008099541 @default.
- W3100809954 hasRelatedWork W1490470283 @default.
- W3100809954 hasRelatedWork W1993233156 @default.
- W3100809954 hasRelatedWork W2105645688 @default.
- W3100809954 hasRelatedWork W2166510624 @default.
- W3100809954 hasRelatedWork W2746080558 @default.
- W3100809954 hasRelatedWork W2889677830 @default.
- W3100809954 hasRelatedWork W3096161060 @default.
- W3100809954 hasRelatedWork W3100809954 @default.
- W3100809954 hasRelatedWork W3132987545 @default.
- W3100809954 hasRelatedWork W4289543958 @default.
- W3100809954 isParatext "false" @default.
- W3100809954 isRetracted "false" @default.
- W3100809954 magId "3100809954" @default.
- W3100809954 workType "article" @default.