Matches in SemOpenAlex for { <https://semopenalex.org/work/W3177172118> ?p ?o ?g. }
- W3177172118 abstract "Fully non-autoregressive neural machine translation (NAT) is proposed to simultaneously predict tokens with single forward of neural networks, which significantly reduces the inference latency at the expense of quality drop compared to the Transformer baseline. In this work, we target on closing the performance gap while maintaining the latency advantage. We first inspect the fundamental issues of fully NAT models, and adopt dependency reduction in the learning space of output tokens as the basic guidance. Then, we revisit methods in four different aspects that have been proven effective for improving NAT models, and carefully combine these techniques with necessary modifications. Our extensive experiments on three translation benchmarks show that the proposed system achieves the new state-of-the-art results for fully NAT models, and obtains comparable performance with the autoregressive and iterative NAT systems. For instance, one of the proposed models achieves 27.49 BLEU points on WMT14 En-De with approximately 16.5X speed up at inference time." @default.
- W3177172118 created "2021-07-05" @default.
- W3177172118 creator A5015250671 @default.
- W3177172118 creator A5076914759 @default.
- W3177172118 date "2021-01-01" @default.
- W3177172118 modified "2023-09-30" @default.
- W3177172118 title "Fully Non-autoregressive Neural Machine Translation: Tricks of the Trade" @default.
- W3177172118 cites W1647671624 @default.
- W3177172118 cites W1821462560 @default.
- W3177172118 cites W2089468765 @default.
- W3177172118 cites W2101105183 @default.
- W3177172118 cites W2127141656 @default.
- W3177172118 cites W2134800885 @default.
- W3177172118 cites W2560512785 @default.
- W3177172118 cites W2805493160 @default.
- W3177172118 cites W2890501761 @default.
- W3177172118 cites W2892213699 @default.
- W3177172118 cites W2920538220 @default.
- W3177172118 cites W2933138175 @default.
- W3177172118 cites W2948381505 @default.
- W3177172118 cites W2948629866 @default.
- W3177172118 cites W2963250244 @default.
- W3177172118 cites W2963366552 @default.
- W3177172118 cites W2963403868 @default.
- W3177172118 cites W2963434219 @default.
- W3177172118 cites W2963532001 @default.
- W3177172118 cites W2963536265 @default.
- W3177172118 cites W2963643655 @default.
- W3177172118 cites W2963946353 @default.
- W3177172118 cites W2964308564 @default.
- W3177172118 cites W2971167892 @default.
- W3177172118 cites W2988975212 @default.
- W3177172118 cites W2990372437 @default.
- W3177172118 cites W2990389671 @default.
- W3177172118 cites W2995999067 @default.
- W3177172118 cites W2996286887 @default.
- W3177172118 cites W2996843693 @default.
- W3177172118 cites W3000840023 @default.
- W3177172118 cites W3015205052 @default.
- W3177172118 cites W3034216012 @default.
- W3177172118 cites W3034363136 @default.
- W3177172118 cites W3034425996 @default.
- W3177172118 cites W3034892578 @default.
- W3177172118 cites W3035416964 @default.
- W3177172118 cites W3035812575 @default.
- W3177172118 cites W3045818398 @default.
- W3177172118 cites W3054488230 @default.
- W3177172118 cites W3099665409 @default.
- W3177172118 cites W3100753857 @default.
- W3177172118 cites W3114869109 @default.
- W3177172118 cites W3127901106 @default.
- W3177172118 cites W3175164646 @default.
- W3177172118 doi "https://doi.org/10.18653/v1/2021.findings-acl.11" @default.
- W3177172118 hasPublicationYear "2021" @default.
- W3177172118 type Work @default.
- W3177172118 sameAs 3177172118 @default.
- W3177172118 citedByCount "12" @default.
- W3177172118 countsByYear W31771721182021 @default.
- W3177172118 countsByYear W31771721182022 @default.
- W3177172118 countsByYear W31771721182023 @default.
- W3177172118 crossrefType "proceedings-article" @default.
- W3177172118 hasAuthorship W3177172118A5015250671 @default.
- W3177172118 hasAuthorship W3177172118A5076914759 @default.
- W3177172118 hasBestOaLocation W31771721181 @default.
- W3177172118 hasConcept C104317684 @default.
- W3177172118 hasConcept C105580179 @default.
- W3177172118 hasConcept C119857082 @default.
- W3177172118 hasConcept C149364088 @default.
- W3177172118 hasConcept C149782125 @default.
- W3177172118 hasConcept C154945302 @default.
- W3177172118 hasConcept C159877910 @default.
- W3177172118 hasConcept C162324750 @default.
- W3177172118 hasConcept C185592680 @default.
- W3177172118 hasConcept C203005215 @default.
- W3177172118 hasConcept C41008148 @default.
- W3177172118 hasConcept C50644808 @default.
- W3177172118 hasConcept C55493867 @default.
- W3177172118 hasConceptScore W3177172118C104317684 @default.
- W3177172118 hasConceptScore W3177172118C105580179 @default.
- W3177172118 hasConceptScore W3177172118C119857082 @default.
- W3177172118 hasConceptScore W3177172118C149364088 @default.
- W3177172118 hasConceptScore W3177172118C149782125 @default.
- W3177172118 hasConceptScore W3177172118C154945302 @default.
- W3177172118 hasConceptScore W3177172118C159877910 @default.
- W3177172118 hasConceptScore W3177172118C162324750 @default.
- W3177172118 hasConceptScore W3177172118C185592680 @default.
- W3177172118 hasConceptScore W3177172118C203005215 @default.
- W3177172118 hasConceptScore W3177172118C41008148 @default.
- W3177172118 hasConceptScore W3177172118C50644808 @default.
- W3177172118 hasConceptScore W3177172118C55493867 @default.
- W3177172118 hasLocation W31771721181 @default.
- W3177172118 hasLocation W31771721182 @default.
- W3177172118 hasOpenAccess W3177172118 @default.
- W3177172118 hasPrimaryLocation W31771721181 @default.
- W3177172118 hasRelatedWork W1974878518 @default.
- W3177172118 hasRelatedWork W2016948484 @default.
- W3177172118 hasRelatedWork W2392593357 @default.
- W3177172118 hasRelatedWork W2435130738 @default.
- W3177172118 hasRelatedWork W2890501761 @default.
- W3177172118 hasRelatedWork W2961085424 @default.