Matches in SemOpenAlex for { <https://semopenalex.org/work/W2891924676> ?p ?o ?g. }
Showing items 1 to 88 of
88
with 100 items per page.
- W2891924676 abstract "Due to the benefits of model compactness, multilingual translation (including many-to-one, many-to-many and one-to-many) based on a universal encoder-decoder architecture attracts more and more attention. However, previous studies show that one-to-many translation based on this framework cannot perform on par with the individually trained models. In this work, we introduce three strategies to improve one-to-many multilingual translation by balancing the shared and unique features. Within the architecture of one decoder for all target languages, we first exploit the use of unique initial states for different target languages. Then, we employ language-dependent positional embeddings. Finally and especially, we propose to divide the hidden cells of the decoder into shared and language-dependent ones. The extensive experiments demonstrate that our proposed methods can obtain remarkable improvements over the strong baselines. Moreover, our strategies can achieve comparable or even better performance than the individually trained translation models." @default.
- W2891924676 created "2018-09-27" @default.
- W2891924676 creator A5001541516 @default.
- W2891924676 creator A5015785439 @default.
- W2891924676 creator A5019899786 @default.
- W2891924676 creator A5024810143 @default.
- W2891924676 creator A5066109863 @default.
- W2891924676 date "2018-01-01" @default.
- W2891924676 modified "2023-09-25" @default.
- W2891924676 title "Three Strategies to Improve One-to-Many Multilingual Translation" @default.
- W2891924676 cites W2101105183 @default.
- W2891924676 cites W2121879602 @default.
- W2891924676 cites W2130942839 @default.
- W2891924676 cites W2251743902 @default.
- W2891924676 cites W2413436069 @default.
- W2891924676 cites W2525778437 @default.
- W2891924676 cites W2539201987 @default.
- W2891924676 cites W2550821151 @default.
- W2891924676 cites W2555745756 @default.
- W2891924676 cites W2561274697 @default.
- W2891924676 cites W2610245951 @default.
- W2891924676 cites W2803258077 @default.
- W2891924676 cites W2962700074 @default.
- W2891924676 cites W2962778428 @default.
- W2891924676 cites W2962807144 @default.
- W2891924676 cites W2963088995 @default.
- W2891924676 cites W2963216553 @default.
- W2891924676 cites W2963247703 @default.
- W2891924676 cites W2963403868 @default.
- W2891924676 cites W2963499433 @default.
- W2891924676 cites W2963842982 @default.
- W2891924676 cites W2963879527 @default.
- W2891924676 cites W2963993537 @default.
- W2891924676 cites W2964007535 @default.
- W2891924676 cites W2964034111 @default.
- W2891924676 cites W2964121744 @default.
- W2891924676 cites W2964265128 @default.
- W2891924676 cites W2964308564 @default.
- W2891924676 doi "https://doi.org/10.18653/v1/d18-1326" @default.
- W2891924676 hasPublicationYear "2018" @default.
- W2891924676 type Work @default.
- W2891924676 sameAs 2891924676 @default.
- W2891924676 citedByCount "69" @default.
- W2891924676 countsByYear W28919246762019 @default.
- W2891924676 countsByYear W28919246762020 @default.
- W2891924676 countsByYear W28919246762021 @default.
- W2891924676 countsByYear W28919246762022 @default.
- W2891924676 countsByYear W28919246762023 @default.
- W2891924676 crossrefType "proceedings-article" @default.
- W2891924676 hasAuthorship W2891924676A5001541516 @default.
- W2891924676 hasAuthorship W2891924676A5015785439 @default.
- W2891924676 hasAuthorship W2891924676A5019899786 @default.
- W2891924676 hasAuthorship W2891924676A5024810143 @default.
- W2891924676 hasAuthorship W2891924676A5066109863 @default.
- W2891924676 hasBestOaLocation W28919246761 @default.
- W2891924676 hasConcept C104317684 @default.
- W2891924676 hasConcept C105580179 @default.
- W2891924676 hasConcept C149364088 @default.
- W2891924676 hasConcept C154945302 @default.
- W2891924676 hasConcept C185592680 @default.
- W2891924676 hasConcept C204321447 @default.
- W2891924676 hasConcept C41008148 @default.
- W2891924676 hasConcept C55493867 @default.
- W2891924676 hasConceptScore W2891924676C104317684 @default.
- W2891924676 hasConceptScore W2891924676C105580179 @default.
- W2891924676 hasConceptScore W2891924676C149364088 @default.
- W2891924676 hasConceptScore W2891924676C154945302 @default.
- W2891924676 hasConceptScore W2891924676C185592680 @default.
- W2891924676 hasConceptScore W2891924676C204321447 @default.
- W2891924676 hasConceptScore W2891924676C41008148 @default.
- W2891924676 hasConceptScore W2891924676C55493867 @default.
- W2891924676 hasLocation W28919246761 @default.
- W2891924676 hasOpenAccess W2891924676 @default.
- W2891924676 hasPrimaryLocation W28919246761 @default.
- W2891924676 hasRelatedWork W1515542156 @default.
- W2891924676 hasRelatedWork W1552159754 @default.
- W2891924676 hasRelatedWork W2131420137 @default.
- W2891924676 hasRelatedWork W2148757832 @default.
- W2891924676 hasRelatedWork W2293457016 @default.
- W2891924676 hasRelatedWork W2368651715 @default.
- W2891924676 hasRelatedWork W2611614995 @default.
- W2891924676 hasRelatedWork W2789919619 @default.
- W2891924676 hasRelatedWork W3107474891 @default.
- W2891924676 hasRelatedWork W3169305685 @default.
- W2891924676 isParatext "false" @default.
- W2891924676 isRetracted "false" @default.
- W2891924676 magId "2891924676" @default.
- W2891924676 workType "article" @default.