FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3211473596> ?p ?o ?g. }

• W3211473596 abstract "Neural machine translation (NMT) has arguably achieved human level parity when trained and evaluated at the sentence-level. Document-level neural machine translation has received less attention and lags behind its sentence-level counterpart. The majority of the proposed document-level approaches investigate ways of conditioning the model on several source or target sentences to capture document context. These approaches require training a specialized NMT model from scratch on parallel document-level corpora. We propose an approach that doesn't require training a specialized model on parallel document-level corpora and is applied to a trained sentence-level NMT model at decoding time. We process the document from left to right multiple times and self-train the sentence-level model on pairs of source sentences and generated translations. Our approach reinforces the choices made by the model, thus making it more likely that the same choices will be made in other sentences in the document. We evaluate our approach on three document-level datasets: NIST Chinese-English, WMT'19 Chinese-English and OpenSubtitles English-Russian. We demonstrate that our approach has higher BLEU score and higher human preference than the baseline. Qualitative analysis of our approach shows that choices made by model are consistent across the document." @default.
• W3211473596 created "2021-11-22" @default.
• W3211473596 creator A5039107666 @default.
• W3211473596 creator A5045225880 @default.
• W3211473596 creator A5067222034 @default.
• W3211473596 date "2020-03-11" @default.
• W3211473596 modified "2023-09-24" @default.
• W3211473596 title "Capturing document context inside sentence-level neural machine translation models with self-training" @default.
• W3211473596 cites W1753482797 @default.
• W3211473596 cites W1810943226 @default.
• W3211473596 cites W1905522558 @default.
• W3211473596 cites W1975879668 @default.
• W3211473596 cites W1995875735 @default.
• W3211473596 cites W2095705004 @default.
• W3211473596 cites W2101210369 @default.
• W3211473596 cites W2111316763 @default.
• W3211473596 cites W2128084896 @default.
• W3211473596 cites W2129217160 @default.
• W3211473596 cites W2130942839 @default.
• W3211473596 cites W2145837098 @default.
• W3211473596 cites W2163568299 @default.
• W3211473596 cites W2535697732 @default.
• W3211473596 cites W2561274697 @default.
• W3211473596 cites W2567571499 @default.
• W3211473596 cites W2623559126 @default.
• W3211473596 cites W2752047430 @default.
• W3211473596 cites W2794365787 @default.
• W3211473596 cites W2806987872 @default.
• W3211473596 cites W2808508619 @default.
• W3211473596 cites W2889326796 @default.
• W3211473596 cites W2891534142 @default.
• W3211473596 cites W2952446148 @default.
• W3211473596 cites W2962712961 @default.
• W3211473596 cites W2962784628 @default.
• W3211473596 cites W2962802109 @default.
• W3211473596 cites W2963216553 @default.
• W3211473596 cites W2963403868 @default.
• W3211473596 cites W2963532001 @default.
• W3211473596 cites W2963833497 @default.
• W3211473596 cites W2963897095 @default.
• W3211473596 cites W2963938518 @default.
• W3211473596 cites W2964093087 @default.
• W3211473596 cites W2964121744 @default.
• W3211473596 cites W2964202354 @default.
• W3211473596 cites W2964289193 @default.
• W3211473596 cites W2964308564 @default.
• W3211473596 cites W2970529093 @default.
• W3211473596 cites W2970694516 @default.
• W3211473596 cites W2970752831 @default.
• W3211473596 cites W2971347700 @default.
• W3211473596 cites W2982644924 @default.
• W3211473596 cites W2983590910 @default.
• W3211473596 cites W2995746049 @default.
• W3211473596 cites W3042199843 @default.
• W3211473596 cites W3120929527 @default.
• W3211473596 cites W44695385 @default.
• W3211473596 cites W780682569 @default.
• W3211473596 cites W85840493 @default.
• W3211473596 doi "https://doi.org/10.48550/arxiv.2003.05259" @default.
• W3211473596 hasPublicationYear "2020" @default.
• W3211473596 type Work @default.
• W3211473596 sameAs 3211473596 @default.
• W3211473596 citedByCount "0" @default.
• W3211473596 crossrefType "posted-content" @default.
• W3211473596 hasAuthorship W3211473596A5039107666 @default.
• W3211473596 hasAuthorship W3211473596A5045225880 @default.
• W3211473596 hasAuthorship W3211473596A5067222034 @default.
• W3211473596 hasBestOaLocation W32114735961 @default.
• W3211473596 hasConcept C111219384 @default.
• W3211473596 hasConcept C135784402 @default.
• W3211473596 hasConcept C148526163 @default.
• W3211473596 hasConcept C151730666 @default.
• W3211473596 hasConcept C154945302 @default.
• W3211473596 hasConcept C162324750 @default.
• W3211473596 hasConcept C175444787 @default.
• W3211473596 hasConcept C203005215 @default.
• W3211473596 hasConcept C204321447 @default.
• W3211473596 hasConcept C24687705 @default.
• W3211473596 hasConcept C2777530160 @default.
• W3211473596 hasConcept C2779343474 @default.
• W3211473596 hasConcept C2781249084 @default.
• W3211473596 hasConcept C28490314 @default.
• W3211473596 hasConcept C41008148 @default.
• W3211473596 hasConcept C86803240 @default.
• W3211473596 hasConceptScore W3211473596C111219384 @default.
• W3211473596 hasConceptScore W3211473596C135784402 @default.
• W3211473596 hasConceptScore W3211473596C148526163 @default.
• W3211473596 hasConceptScore W3211473596C151730666 @default.
• W3211473596 hasConceptScore W3211473596C154945302 @default.
• W3211473596 hasConceptScore W3211473596C162324750 @default.
• W3211473596 hasConceptScore W3211473596C175444787 @default.
• W3211473596 hasConceptScore W3211473596C203005215 @default.
• W3211473596 hasConceptScore W3211473596C204321447 @default.
• W3211473596 hasConceptScore W3211473596C24687705 @default.
• W3211473596 hasConceptScore W3211473596C2777530160 @default.
• W3211473596 hasConceptScore W3211473596C2779343474 @default.
• W3211473596 hasConceptScore W3211473596C2781249084 @default.
• W3211473596 hasConceptScore W3211473596C28490314 @default.
• W3211473596 hasConceptScore W3211473596C41008148 @default.
• W3211473596 hasConceptScore W3211473596C86803240 @default.

• next