FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3176466730> ?p ?o ?g. }

• W3176466730 abstract "The data scarcity in low-resource languages has become a bottleneck to building robust neural machine translation systems. Fine-tuning a multilingual pre-trained model (e.g., mBART (Liu et al., 2020)) on the translation task is a good approach for low-resource languages; however, its performance will be greatly limited when there are unseen languages in the translation pairs. In this paper, we present a continual pre-training (CPT) framework on mBART to effectively adapt it to unseen languages. We first construct noisy mixed-language text from the monolingual corpus of the target language in the translation pair to cover both the source and target languages, and then, we continue pre-training mBART to reconstruct the original monolingual text. Results show that our method can consistently improve the fine-tuning performance upon the mBART baseline, as well as other strong baselines, across all tested low-resource translation pairs containing unseen languages. Furthermore, our approach also boosts the performance on translation pairs where both languages are seen in the original mBART's pre-training. The code is available at this https URL." @default.
• W3176466730 created "2021-07-05" @default.
• W3176466730 creator A5011760791 @default.
• W3176466730 creator A5065856469 @default.
• W3176466730 creator A5085516032 @default.
• W3176466730 date "2021-01-01" @default.
• W3176466730 modified "2023-10-17" @default.
• W3176466730 title "Continual Mixed-Language Pre-Training for Extremely Low-Resource Neural Machine Translation" @default.
• W3176466730 cites W1508577659 @default.
• W3176466730 cites W1557247526 @default.
• W3176466730 cites W1902237438 @default.
• W3176466730 cites W2101105183 @default.
• W3176466730 cites W2419539795 @default.
• W3176466730 cites W2756566411 @default.
• W3176466730 cites W2786253471 @default.
• W3176466730 cites W2803172920 @default.
• W3176466730 cites W2913659301 @default.
• W3176466730 cites W2944815030 @default.
• W3176466730 cites W2951476960 @default.
• W3176466730 cites W2962824887 @default.
• W3176466730 cites W2963088995 @default.
• W3176466730 cites W2963216553 @default.
• W3176466730 cites W2963341956 @default.
• W3176466730 cites W2963403868 @default.
• W3176466730 cites W2963602293 @default.
• W3176466730 cites W2963741406 @default.
• W3176466730 cites W2963993537 @default.
• W3176466730 cites W2964121744 @default.
• W3176466730 cites W2964308564 @default.
• W3176466730 cites W2970876710 @default.
• W3176466730 cites W2971167006 @default.
• W3176466730 cites W2971863715 @default.
• W3176466730 cites W2986030184 @default.
• W3176466730 cites W2997214274 @default.
• W3176466730 cites W3035144493 @default.
• W3176466730 cites W3035390927 @default.
• W3176466730 cites W3035565536 @default.
• W3176466730 cites W3046368065 @default.
• W3176466730 cites W3082274269 @default.
• W3176466730 cites W3088382025 @default.
• W3176466730 cites W3092327118 @default.
• W3176466730 cites W3093871477 @default.
• W3176466730 cites W3098341425 @default.
• W3176466730 cites W3098466758 @default.
• W3176466730 cites W3100198908 @default.
• W3176466730 cites W3102854726 @default.
• W3176466730 cites W3104273515 @default.
• W3176466730 cites W3104652516 @default.
• W3176466730 cites W3105912780 @default.
• W3176466730 cites W3107826490 @default.
• W3176466730 cites W3127887696 @default.
• W3176466730 cites W3153805297 @default.
• W3176466730 cites W3169483174 @default.
• W3176466730 cites W3169565655 @default.
• W3176466730 cites W3174366011 @default.
• W3176466730 doi "https://doi.org/10.18653/v1/2021.findings-acl.239" @default.
• W3176466730 hasPublicationYear "2021" @default.
• W3176466730 type Work @default.
• W3176466730 sameAs 3176466730 @default.
• W3176466730 citedByCount "3" @default.
• W3176466730 countsByYear W31764667302021 @default.
• W3176466730 countsByYear W31764667302023 @default.
• W3176466730 crossrefType "proceedings-article" @default.
• W3176466730 hasAuthorship W3176466730A5011760791 @default.
• W3176466730 hasAuthorship W3176466730A5065856469 @default.
• W3176466730 hasAuthorship W3176466730A5085516032 @default.
• W3176466730 hasBestOaLocation W31764667301 @default.
• W3176466730 hasConcept C104317684 @default.
• W3176466730 hasConcept C105580179 @default.
• W3176466730 hasConcept C121332964 @default.
• W3176466730 hasConcept C149364088 @default.
• W3176466730 hasConcept C153294291 @default.
• W3176466730 hasConcept C154945302 @default.
• W3176466730 hasConcept C185592680 @default.
• W3176466730 hasConcept C203005215 @default.
• W3176466730 hasConcept C204321447 @default.
• W3176466730 hasConcept C2777211547 @default.
• W3176466730 hasConcept C41008148 @default.
• W3176466730 hasConcept C55493867 @default.
• W3176466730 hasConceptScore W3176466730C104317684 @default.
• W3176466730 hasConceptScore W3176466730C105580179 @default.
• W3176466730 hasConceptScore W3176466730C121332964 @default.
• W3176466730 hasConceptScore W3176466730C149364088 @default.
• W3176466730 hasConceptScore W3176466730C153294291 @default.
• W3176466730 hasConceptScore W3176466730C154945302 @default.
• W3176466730 hasConceptScore W3176466730C185592680 @default.
• W3176466730 hasConceptScore W3176466730C203005215 @default.
• W3176466730 hasConceptScore W3176466730C204321447 @default.
• W3176466730 hasConceptScore W3176466730C2777211547 @default.
• W3176466730 hasConceptScore W3176466730C41008148 @default.
• W3176466730 hasConceptScore W3176466730C55493867 @default.
• W3176466730 hasLocation W31764667301 @default.
• W3176466730 hasLocation W31764667302 @default.
• W3176466730 hasOpenAccess W3176466730 @default.
• W3176466730 hasPrimaryLocation W31764667301 @default.
• W3176466730 hasRelatedWork W1484029852 @default.
• W3176466730 hasRelatedWork W1512718085 @default.
• W3176466730 hasRelatedWork W1585034923 @default.
• W3176466730 hasRelatedWork W2167662847 @default.
• W3176466730 hasRelatedWork W2313858059 @default.

• next