Matches in SemOpenAlex for { <https://semopenalex.org/work/W2993091699> ?p ?o ?g. }
- W2993091699 endingPage "141" @default.
- W2993091699 startingPage "134" @default.
- W2993091699 abstract "Recurrent neural networks (RNNs) have recently achieved remarkable successes in a number of applications. However, the huge sizes and computational burden of these models make it difficult for their deployment on edge devices. A practically effective approach is to reduce the overall storage and computation costs of RNNs by network pruning techniques. Despite their successful applications, those pruning methods based on Lasso either produce irregular sparse patterns in weight matrices, which is not helpful in practical speedup. To address these issues, we propose a structured pruning method through neuron selection which can remove the independent neuron of RNNs. More specifically, we introduce two sets of binary random variables, which can be interpreted as gates or switches to the input neurons and the hidden neurons, respectively. We demonstrate that the corresponding optimization problem can be addressed by minimizing the L0 norm of the weight matrix. Finally, experimental results on language modeling and machine reading comprehension tasks have indicated the advantages of the proposed method in comparison with state-of-the-art pruning competitors. In particular, nearly 20× practical speedup during inference was achieved without losing performance for the language model on the Penn TreeBank dataset, indicating the promising performance of the proposed method." @default.
- W2993091699 created "2019-12-13" @default.
- W2993091699 creator A5010510537 @default.
- W2993091699 creator A5041793389 @default.
- W2993091699 creator A5051227924 @default.
- W2993091699 creator A5080943673 @default.
- W2993091699 date "2020-03-01" @default.
- W2993091699 modified "2023-10-17" @default.
- W2993091699 title "Structured pruning of recurrent neural networks through neuron selection" @default.
- W2993091699 cites W1902237438 @default.
- W2993091699 cites W1909234690 @default.
- W2993091699 cites W1999974018 @default.
- W2993091699 cites W2064675550 @default.
- W2993091699 cites W2083508718 @default.
- W2993091699 cites W2119717200 @default.
- W2993091699 cites W2463955103 @default.
- W2993091699 cites W2473409175 @default.
- W2993091699 cites W2546487983 @default.
- W2993091699 cites W2562749854 @default.
- W2993091699 cites W2605766634 @default.
- W2993091699 cites W2745461083 @default.
- W2993091699 cites W2803834712 @default.
- W2993091699 cites W2807812398 @default.
- W2993091699 cites W2889580470 @default.
- W2993091699 cites W2928560789 @default.
- W2993091699 cites W2943925420 @default.
- W2993091699 cites W2963838731 @default.
- W2993091699 cites W2966234499 @default.
- W2993091699 cites W2976540144 @default.
- W2993091699 cites W566555209 @default.
- W2993091699 cites W2293082389 @default.
- W2993091699 doi "https://doi.org/10.1016/j.neunet.2019.11.018" @default.
- W2993091699 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/31855748" @default.
- W2993091699 hasPublicationYear "2020" @default.
- W2993091699 type Work @default.
- W2993091699 sameAs 2993091699 @default.
- W2993091699 citedByCount "24" @default.
- W2993091699 countsByYear W29930916992020 @default.
- W2993091699 countsByYear W29930916992021 @default.
- W2993091699 countsByYear W29930916992022 @default.
- W2993091699 countsByYear W29930916992023 @default.
- W2993091699 crossrefType "journal-article" @default.
- W2993091699 hasAuthorship W2993091699A5010510537 @default.
- W2993091699 hasAuthorship W2993091699A5041793389 @default.
- W2993091699 hasAuthorship W2993091699A5051227924 @default.
- W2993091699 hasAuthorship W2993091699A5080943673 @default.
- W2993091699 hasBestOaLocation W29930916992 @default.
- W2993091699 hasConcept C108010975 @default.
- W2993091699 hasConcept C11413529 @default.
- W2993091699 hasConcept C119857082 @default.
- W2993091699 hasConcept C137293760 @default.
- W2993091699 hasConcept C147168706 @default.
- W2993091699 hasConcept C154945302 @default.
- W2993091699 hasConcept C173608175 @default.
- W2993091699 hasConcept C186644900 @default.
- W2993091699 hasConcept C206134035 @default.
- W2993091699 hasConcept C2776214188 @default.
- W2993091699 hasConcept C41008148 @default.
- W2993091699 hasConcept C45374587 @default.
- W2993091699 hasConcept C50644808 @default.
- W2993091699 hasConcept C6557445 @default.
- W2993091699 hasConcept C68339613 @default.
- W2993091699 hasConcept C86803240 @default.
- W2993091699 hasConceptScore W2993091699C108010975 @default.
- W2993091699 hasConceptScore W2993091699C11413529 @default.
- W2993091699 hasConceptScore W2993091699C119857082 @default.
- W2993091699 hasConceptScore W2993091699C137293760 @default.
- W2993091699 hasConceptScore W2993091699C147168706 @default.
- W2993091699 hasConceptScore W2993091699C154945302 @default.
- W2993091699 hasConceptScore W2993091699C173608175 @default.
- W2993091699 hasConceptScore W2993091699C186644900 @default.
- W2993091699 hasConceptScore W2993091699C206134035 @default.
- W2993091699 hasConceptScore W2993091699C2776214188 @default.
- W2993091699 hasConceptScore W2993091699C41008148 @default.
- W2993091699 hasConceptScore W2993091699C45374587 @default.
- W2993091699 hasConceptScore W2993091699C50644808 @default.
- W2993091699 hasConceptScore W2993091699C6557445 @default.
- W2993091699 hasConceptScore W2993091699C68339613 @default.
- W2993091699 hasConceptScore W2993091699C86803240 @default.
- W2993091699 hasFunder F4320321001 @default.
- W2993091699 hasFunder F4320335787 @default.
- W2993091699 hasLocation W29930916991 @default.
- W2993091699 hasLocation W29930916992 @default.
- W2993091699 hasLocation W29930916993 @default.
- W2993091699 hasOpenAccess W2993091699 @default.
- W2993091699 hasPrimaryLocation W29930916991 @default.
- W2993091699 hasRelatedWork W108437174 @default.
- W2993091699 hasRelatedWork W2747579829 @default.
- W2993091699 hasRelatedWork W2754526845 @default.
- W2993091699 hasRelatedWork W2901663942 @default.
- W2993091699 hasRelatedWork W2952423590 @default.
- W2993091699 hasRelatedWork W2962883174 @default.
- W2993091699 hasRelatedWork W2993091699 @default.
- W2993091699 hasRelatedWork W3168100846 @default.
- W2993091699 hasRelatedWork W3196045513 @default.
- W2993091699 hasRelatedWork W4287021266 @default.
- W2993091699 hasVolume "123" @default.
- W2993091699 isParatext "false" @default.