Matches in SemOpenAlex for { <https://semopenalex.org/work/W2948381505> ?p ?o ?g. }
- W2948381505 abstract "Non-autoregressive translation models (NAT) have achieved impressive inference speedup. A potential issue of the existing NAT algorithms, however, is that the decoding is conducted in parallel, without directly considering previous context. In this paper, we propose an imitation learning framework for non-autoregressive machine translation, which still enjoys the fast translation speed but gives comparable translation performance compared to its auto-regressive counterpart. We conduct experiments on the IWSLT16, WMT14 and WMT16 datasets. Our proposed model achieves a significant speedup over the autoregressive models, while keeping the translation quality comparable to the autoregressive models. By sampling sentence length in parallel at inference time, we achieve the performance of 31.85 BLEU on WMT16 Ro$rightarrow$En and 30.68 BLEU on IWSLT16 En$rightarrow$De." @default.
- W2948381505 created "2019-06-14" @default.
- W2948381505 creator A5000258378 @default.
- W2948381505 creator A5006367598 @default.
- W2948381505 creator A5006497264 @default.
- W2948381505 creator A5022574607 @default.
- W2948381505 creator A5043256966 @default.
- W2948381505 creator A5057519041 @default.
- W2948381505 date "2019-06-05" @default.
- W2948381505 modified "2023-09-23" @default.
- W2948381505 title "Imitation Learning for Non-Autoregressive Neural Machine Translation" @default.
- W2948381505 cites W1757796397 @default.
- W2948381505 cites W1916559533 @default.
- W2948381505 cites W2064675550 @default.
- W2948381505 cites W2072128103 @default.
- W2948381505 cites W2124807415 @default.
- W2948381505 cites W2130942839 @default.
- W2948381505 cites W2145339207 @default.
- W2948381505 cites W2153653739 @default.
- W2948381505 cites W2157331557 @default.
- W2948381505 cites W2463507112 @default.
- W2948381505 cites W2518578398 @default.
- W2948381505 cites W2626778328 @default.
- W2948381505 cites W2767206889 @default.
- W2948381505 cites W2789543585 @default.
- W2948381505 cites W2892213699 @default.
- W2948381505 cites W2903592137 @default.
- W2948381505 cites W2903918613 @default.
- W2948381505 cites W2962784628 @default.
- W2948381505 cites W2963434219 @default.
- W2948381505 cites W2964074409 @default.
- W2948381505 cites W3141754740 @default.
- W2948381505 doi "https://doi.org/10.48550/arxiv.1906.02041" @default.
- W2948381505 hasPublicationYear "2019" @default.
- W2948381505 type Work @default.
- W2948381505 sameAs 2948381505 @default.
- W2948381505 citedByCount "21" @default.
- W2948381505 countsByYear W29483815052019 @default.
- W2948381505 countsByYear W29483815052020 @default.
- W2948381505 countsByYear W29483815052021 @default.
- W2948381505 countsByYear W29483815052022 @default.
- W2948381505 crossrefType "posted-content" @default.
- W2948381505 hasAuthorship W2948381505A5000258378 @default.
- W2948381505 hasAuthorship W2948381505A5006367598 @default.
- W2948381505 hasAuthorship W2948381505A5006497264 @default.
- W2948381505 hasAuthorship W2948381505A5022574607 @default.
- W2948381505 hasAuthorship W2948381505A5043256966 @default.
- W2948381505 hasAuthorship W2948381505A5057519041 @default.
- W2948381505 hasBestOaLocation W29483815051 @default.
- W2948381505 hasConcept C104317684 @default.
- W2948381505 hasConcept C105580179 @default.
- W2948381505 hasConcept C119857082 @default.
- W2948381505 hasConcept C126388530 @default.
- W2948381505 hasConcept C149364088 @default.
- W2948381505 hasConcept C149782125 @default.
- W2948381505 hasConcept C151730666 @default.
- W2948381505 hasConcept C154945302 @default.
- W2948381505 hasConcept C15744967 @default.
- W2948381505 hasConcept C159877910 @default.
- W2948381505 hasConcept C173608175 @default.
- W2948381505 hasConcept C185592680 @default.
- W2948381505 hasConcept C203005215 @default.
- W2948381505 hasConcept C2776214188 @default.
- W2948381505 hasConcept C2779343474 @default.
- W2948381505 hasConcept C28490314 @default.
- W2948381505 hasConcept C33923547 @default.
- W2948381505 hasConcept C41008148 @default.
- W2948381505 hasConcept C55493867 @default.
- W2948381505 hasConcept C68339613 @default.
- W2948381505 hasConcept C77805123 @default.
- W2948381505 hasConcept C86803240 @default.
- W2948381505 hasConceptScore W2948381505C104317684 @default.
- W2948381505 hasConceptScore W2948381505C105580179 @default.
- W2948381505 hasConceptScore W2948381505C119857082 @default.
- W2948381505 hasConceptScore W2948381505C126388530 @default.
- W2948381505 hasConceptScore W2948381505C149364088 @default.
- W2948381505 hasConceptScore W2948381505C149782125 @default.
- W2948381505 hasConceptScore W2948381505C151730666 @default.
- W2948381505 hasConceptScore W2948381505C154945302 @default.
- W2948381505 hasConceptScore W2948381505C15744967 @default.
- W2948381505 hasConceptScore W2948381505C159877910 @default.
- W2948381505 hasConceptScore W2948381505C173608175 @default.
- W2948381505 hasConceptScore W2948381505C185592680 @default.
- W2948381505 hasConceptScore W2948381505C203005215 @default.
- W2948381505 hasConceptScore W2948381505C2776214188 @default.
- W2948381505 hasConceptScore W2948381505C2779343474 @default.
- W2948381505 hasConceptScore W2948381505C28490314 @default.
- W2948381505 hasConceptScore W2948381505C33923547 @default.
- W2948381505 hasConceptScore W2948381505C41008148 @default.
- W2948381505 hasConceptScore W2948381505C55493867 @default.
- W2948381505 hasConceptScore W2948381505C68339613 @default.
- W2948381505 hasConceptScore W2948381505C77805123 @default.
- W2948381505 hasConceptScore W2948381505C86803240 @default.
- W2948381505 hasLocation W29483815051 @default.
- W2948381505 hasOpenAccess W2948381505 @default.
- W2948381505 hasPrimaryLocation W29483815051 @default.
- W2948381505 hasRelatedWork W2948381505 @default.
- W2948381505 hasRelatedWork W2962915948 @default.
- W2948381505 hasRelatedWork W2969565141 @default.
- W2948381505 hasRelatedWork W2990488589 @default.