Matches in SemOpenAlex for { <https://semopenalex.org/work/W2952423590> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W2952423590 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 structured pruning method through neuron selection which can reduce the sizes of basic structures 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 x practical speedup during inference was achieved without losing performance for language model on the Penn TreeBank dataset, indicating the promising performance of the proposed method" @default.
- W2952423590 created "2019-06-27" @default.
- W2952423590 creator A5010510537 @default.
- W2952423590 creator A5041793389 @default.
- W2952423590 creator A5051227924 @default.
- W2952423590 creator A5080943673 @default.
- W2952423590 date "2019-06-17" @default.
- W2952423590 modified "2023-10-14" @default.
- W2952423590 title "Structured Pruning of Recurrent Neural Networks through Neuron Selection" @default.
- W2952423590 cites W1632114991 @default.
- W2952423590 cites W1902237438 @default.
- W2952423590 cites W1909234690 @default.
- W2952423590 cites W1999974018 @default.
- W2952423590 cites W2064675550 @default.
- W2952423590 cites W2083508718 @default.
- W2952423590 cites W2119717200 @default.
- W2952423590 cites W2130942839 @default.
- W2952423590 cites W2312434537 @default.
- W2952423590 cites W2463955103 @default.
- W2952423590 cites W2473409175 @default.
- W2952423590 cites W2546487983 @default.
- W2952423590 cites W2551895583 @default.
- W2952423590 cites W2562749854 @default.
- W2952423590 cites W2605766634 @default.
- W2952423590 cites W2745461083 @default.
- W2952423590 cites W2807812398 @default.
- W2952423590 cites W2889580470 @default.
- W2952423590 cites W2889847962 @default.
- W2952423590 cites W2928560789 @default.
- W2952423590 cites W2943925420 @default.
- W2952423590 cites W2963000224 @default.
- W2952423590 cites W2963094099 @default.
- W2952423590 cites W2963363373 @default.
- W2952423590 cites W2963838731 @default.
- W2952423590 cites W2963983719 @default.
- W2952423590 cites W2966234499 @default.
- W2952423590 cites W2976540144 @default.
- W2952423590 cites W566555209 @default.
- W2952423590 cites W2293082389 @default.
- W2952423590 doi "https://doi.org/10.48550/arxiv.1906.06847" @default.
- W2952423590 hasPublicationYear "2019" @default.
- W2952423590 type Work @default.
- W2952423590 sameAs 2952423590 @default.
- W2952423590 citedByCount "0" @default.
- W2952423590 crossrefType "posted-content" @default.
- W2952423590 hasAuthorship W2952423590A5010510537 @default.
- W2952423590 hasAuthorship W2952423590A5041793389 @default.
- W2952423590 hasAuthorship W2952423590A5051227924 @default.
- W2952423590 hasAuthorship W2952423590A5080943673 @default.
- W2952423590 hasBestOaLocation W29524235901 @default.
- W2952423590 hasConcept C108010975 @default.
- W2952423590 hasConcept C11413529 @default.
- W2952423590 hasConcept C119857082 @default.
- W2952423590 hasConcept C137293760 @default.
- W2952423590 hasConcept C147168706 @default.
- W2952423590 hasConcept C154945302 @default.
- W2952423590 hasConcept C173608175 @default.
- W2952423590 hasConcept C186644900 @default.
- W2952423590 hasConcept C206134035 @default.
- W2952423590 hasConcept C2776214188 @default.
- W2952423590 hasConcept C41008148 @default.
- W2952423590 hasConcept C45374587 @default.
- W2952423590 hasConcept C50644808 @default.
- W2952423590 hasConcept C6557445 @default.
- W2952423590 hasConcept C68339613 @default.
- W2952423590 hasConcept C86803240 @default.
- W2952423590 hasConceptScore W2952423590C108010975 @default.
- W2952423590 hasConceptScore W2952423590C11413529 @default.
- W2952423590 hasConceptScore W2952423590C119857082 @default.
- W2952423590 hasConceptScore W2952423590C137293760 @default.
- W2952423590 hasConceptScore W2952423590C147168706 @default.
- W2952423590 hasConceptScore W2952423590C154945302 @default.
- W2952423590 hasConceptScore W2952423590C173608175 @default.
- W2952423590 hasConceptScore W2952423590C186644900 @default.
- W2952423590 hasConceptScore W2952423590C206134035 @default.
- W2952423590 hasConceptScore W2952423590C2776214188 @default.
- W2952423590 hasConceptScore W2952423590C41008148 @default.
- W2952423590 hasConceptScore W2952423590C45374587 @default.
- W2952423590 hasConceptScore W2952423590C50644808 @default.
- W2952423590 hasConceptScore W2952423590C6557445 @default.
- W2952423590 hasConceptScore W2952423590C68339613 @default.
- W2952423590 hasConceptScore W2952423590C86803240 @default.
- W2952423590 hasLocation W29524235901 @default.
- W2952423590 hasOpenAccess W2952423590 @default.
- W2952423590 hasPrimaryLocation W29524235901 @default.
- W2952423590 hasRelatedWork W108437174 @default.
- W2952423590 hasRelatedWork W2747579829 @default.
- W2952423590 hasRelatedWork W2754526845 @default.
- W2952423590 hasRelatedWork W2755192592 @default.
- W2952423590 hasRelatedWork W2952423590 @default.
- W2952423590 hasRelatedWork W2962883174 @default.
- W2952423590 hasRelatedWork W3168100846 @default.
- W2952423590 hasRelatedWork W3196045513 @default.
- W2952423590 hasRelatedWork W4287021266 @default.
- W2952423590 hasRelatedWork W4289294962 @default.
- W2952423590 isParatext "false" @default.
- W2952423590 isRetracted "false" @default.
- W2952423590 magId "2952423590" @default.
- W2952423590 workType "article" @default.