Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313413720> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4313413720 abstract "We introduce dynamic auto-sizing, a novel approach to training artificial neural networks which allows the models to automatically adapt their size to the problem domain. The size of the models can be further controlled during the learning process by modifying the applied strength of regularization. The ability of dynamic auto-sizing models to expand or shrink their hidden layers is achieved by periodically growing and pruning entire units such as neurons or filters. For this purpose, we introduce weighted L1 regularization, a novel regularization method for inducing structured sparsity. Besides analyzing the behavior of dynamic auto-sizing, we evaluate predictive performance of models trained using the method and show that such models can provide a predictive advantage over traditional approaches." @default.
- W4313413720 created "2023-01-06" @default.
- W4313413720 creator A5035026045 @default.
- W4313413720 creator A5059300694 @default.
- W4313413720 creator A5089137994 @default.
- W4313413720 date "2022-11-10" @default.
- W4313413720 modified "2023-09-27" @default.
- W4313413720 title "Adapting the Size of Artificial Neural Networks Using Dynamic Auto-Sizing" @default.
- W4313413720 cites W1553939018 @default.
- W4313413720 cites W2111935653 @default.
- W4313413720 cites W2119814172 @default.
- W4313413720 cites W2128033389 @default.
- W4313413720 cites W2135046866 @default.
- W4313413720 cites W2138019504 @default.
- W4313413720 cites W2520760693 @default.
- W4313413720 cites W2593744649 @default.
- W4313413720 cites W2964217527 @default.
- W4313413720 doi "https://doi.org/10.1109/csit56902.2022.10000471" @default.
- W4313413720 hasPublicationYear "2022" @default.
- W4313413720 type Work @default.
- W4313413720 citedByCount "0" @default.
- W4313413720 crossrefType "proceedings-article" @default.
- W4313413720 hasAuthorship W4313413720A5035026045 @default.
- W4313413720 hasAuthorship W4313413720A5059300694 @default.
- W4313413720 hasAuthorship W4313413720A5089137994 @default.
- W4313413720 hasConcept C108010975 @default.
- W4313413720 hasConcept C111919701 @default.
- W4313413720 hasConcept C119857082 @default.
- W4313413720 hasConcept C142362112 @default.
- W4313413720 hasConcept C153349607 @default.
- W4313413720 hasConcept C154945302 @default.
- W4313413720 hasConcept C2776135515 @default.
- W4313413720 hasConcept C2777767291 @default.
- W4313413720 hasConcept C41008148 @default.
- W4313413720 hasConcept C50644808 @default.
- W4313413720 hasConcept C6557445 @default.
- W4313413720 hasConcept C86803240 @default.
- W4313413720 hasConcept C98045186 @default.
- W4313413720 hasConceptScore W4313413720C108010975 @default.
- W4313413720 hasConceptScore W4313413720C111919701 @default.
- W4313413720 hasConceptScore W4313413720C119857082 @default.
- W4313413720 hasConceptScore W4313413720C142362112 @default.
- W4313413720 hasConceptScore W4313413720C153349607 @default.
- W4313413720 hasConceptScore W4313413720C154945302 @default.
- W4313413720 hasConceptScore W4313413720C2776135515 @default.
- W4313413720 hasConceptScore W4313413720C2777767291 @default.
- W4313413720 hasConceptScore W4313413720C41008148 @default.
- W4313413720 hasConceptScore W4313413720C50644808 @default.
- W4313413720 hasConceptScore W4313413720C6557445 @default.
- W4313413720 hasConceptScore W4313413720C86803240 @default.
- W4313413720 hasConceptScore W4313413720C98045186 @default.
- W4313413720 hasLocation W43134137201 @default.
- W4313413720 hasOpenAccess W4313413720 @default.
- W4313413720 hasPrimaryLocation W43134137201 @default.
- W4313413720 hasRelatedWork W1982472074 @default.
- W4313413720 hasRelatedWork W2387747922 @default.
- W4313413720 hasRelatedWork W2961085424 @default.
- W4313413720 hasRelatedWork W3046775127 @default.
- W4313413720 hasRelatedWork W3047144510 @default.
- W4313413720 hasRelatedWork W4285260836 @default.
- W4313413720 hasRelatedWork W4286629047 @default.
- W4313413720 hasRelatedWork W4306321456 @default.
- W4313413720 hasRelatedWork W4306674287 @default.
- W4313413720 hasRelatedWork W4224009465 @default.
- W4313413720 isParatext "false" @default.
- W4313413720 isRetracted "false" @default.
- W4313413720 workType "article" @default.