Matches in SemOpenAlex for { <https://semopenalex.org/work/W3207675432> ?p ?o ?g. }
Showing items 1 to 98 of
98
with 100 items per page.
- W3207675432 endingPage "140" @default.
- W3207675432 startingPage "129" @default.
- W3207675432 abstract "Multilingual neural machine translation (MNMT) with a single encoder-decoder model has attracted much interest due to its simple deployment and low training cost. However, the all-shared translation model often yields degraded performance due to the modeling capacity limitations and language diversity. Moreover, it has been revealed in recent studies that the shared parameters lead to negative language interference although they may also facilitate knowledge transfer across languages. In this work, we propose an adaptive architecture for multilingual modeling, which divides the parameters in MNMT sub-layers into shared and language-specific ones. We train the model to learn and balance the shared and unique features with different degrees of parameter sharing. We evaluate our model on one-to-many and many-to-one translation tasks. Experiments on IWSLT dataset show that our proposed model remarkably outperforms the multilingual baseline model and achieves comparable or even better performance compared with the bilingual model." @default.
- W3207675432 created "2021-10-25" @default.
- W3207675432 creator A5013881064 @default.
- W3207675432 creator A5031577422 @default.
- W3207675432 creator A5046871248 @default.
- W3207675432 creator A5049192960 @default.
- W3207675432 creator A5050030131 @default.
- W3207675432 date "2021-01-01" @default.
- W3207675432 modified "2023-10-06" @default.
- W3207675432 title "Adaptive Transformer for Multilingual Neural Machine Translation" @default.
- W3207675432 cites W2251743902 @default.
- W3207675432 cites W2550821151 @default.
- W3207675432 cites W2891924676 @default.
- W3207675432 cites W2899015110 @default.
- W3207675432 cites W2919290281 @default.
- W3207675432 cites W2953190730 @default.
- W3207675432 cites W2962778428 @default.
- W3207675432 cites W2962784628 @default.
- W3207675432 cites W2963088995 @default.
- W3207675432 cites W2963149635 @default.
- W3207675432 cites W2963247703 @default.
- W3207675432 cites W2963532001 @default.
- W3207675432 cites W2963983698 @default.
- W3207675432 cites W2964007535 @default.
- W3207675432 cites W2964034111 @default.
- W3207675432 cites W2970925677 @default.
- W3207675432 cites W3017454464 @default.
- W3207675432 cites W3035390927 @default.
- W3207675432 cites W3090350559 @default.
- W3207675432 cites W3105005398 @default.
- W3207675432 doi "https://doi.org/10.1007/978-3-030-88480-2_11" @default.
- W3207675432 hasPublicationYear "2021" @default.
- W3207675432 type Work @default.
- W3207675432 sameAs 3207675432 @default.
- W3207675432 citedByCount "0" @default.
- W3207675432 crossrefType "book-chapter" @default.
- W3207675432 hasAuthorship W3207675432A5013881064 @default.
- W3207675432 hasAuthorship W3207675432A5031577422 @default.
- W3207675432 hasAuthorship W3207675432A5046871248 @default.
- W3207675432 hasAuthorship W3207675432A5049192960 @default.
- W3207675432 hasAuthorship W3207675432A5050030131 @default.
- W3207675432 hasConcept C104317684 @default.
- W3207675432 hasConcept C105339364 @default.
- W3207675432 hasConcept C105580179 @default.
- W3207675432 hasConcept C111919701 @default.
- W3207675432 hasConcept C115903868 @default.
- W3207675432 hasConcept C118505674 @default.
- W3207675432 hasConcept C119857082 @default.
- W3207675432 hasConcept C121332964 @default.
- W3207675432 hasConcept C137293760 @default.
- W3207675432 hasConcept C149364088 @default.
- W3207675432 hasConcept C154945302 @default.
- W3207675432 hasConcept C165801399 @default.
- W3207675432 hasConcept C185592680 @default.
- W3207675432 hasConcept C203005215 @default.
- W3207675432 hasConcept C204321447 @default.
- W3207675432 hasConcept C41008148 @default.
- W3207675432 hasConcept C55493867 @default.
- W3207675432 hasConcept C62520636 @default.
- W3207675432 hasConcept C66322947 @default.
- W3207675432 hasConceptScore W3207675432C104317684 @default.
- W3207675432 hasConceptScore W3207675432C105339364 @default.
- W3207675432 hasConceptScore W3207675432C105580179 @default.
- W3207675432 hasConceptScore W3207675432C111919701 @default.
- W3207675432 hasConceptScore W3207675432C115903868 @default.
- W3207675432 hasConceptScore W3207675432C118505674 @default.
- W3207675432 hasConceptScore W3207675432C119857082 @default.
- W3207675432 hasConceptScore W3207675432C121332964 @default.
- W3207675432 hasConceptScore W3207675432C137293760 @default.
- W3207675432 hasConceptScore W3207675432C149364088 @default.
- W3207675432 hasConceptScore W3207675432C154945302 @default.
- W3207675432 hasConceptScore W3207675432C165801399 @default.
- W3207675432 hasConceptScore W3207675432C185592680 @default.
- W3207675432 hasConceptScore W3207675432C203005215 @default.
- W3207675432 hasConceptScore W3207675432C204321447 @default.
- W3207675432 hasConceptScore W3207675432C41008148 @default.
- W3207675432 hasConceptScore W3207675432C55493867 @default.
- W3207675432 hasConceptScore W3207675432C62520636 @default.
- W3207675432 hasConceptScore W3207675432C66322947 @default.
- W3207675432 hasLocation W32076754321 @default.
- W3207675432 hasOpenAccess W3207675432 @default.
- W3207675432 hasPrimaryLocation W32076754321 @default.
- W3207675432 hasRelatedWork W2606032440 @default.
- W3207675432 hasRelatedWork W3033942572 @default.
- W3207675432 hasRelatedWork W3107474891 @default.
- W3207675432 hasRelatedWork W3110288483 @default.
- W3207675432 hasRelatedWork W3119899169 @default.
- W3207675432 hasRelatedWork W3204726280 @default.
- W3207675432 hasRelatedWork W3212566403 @default.
- W3207675432 hasRelatedWork W4287761227 @default.
- W3207675432 hasRelatedWork W4288026811 @default.
- W3207675432 hasRelatedWork W61293283 @default.
- W3207675432 isParatext "false" @default.
- W3207675432 isRetracted "false" @default.
- W3207675432 magId "3207675432" @default.
- W3207675432 workType "book-chapter" @default.