Matches in SemOpenAlex for { <https://semopenalex.org/work/W2945501029> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W2945501029 abstract "To improve low-resource Neural Machine Translation (NMT) with multilingual corpora, training on the most related high-resource language only is often more effective than using all data available (Neubig and Hu, 2018). However, it is possible that an intelligent data selection strategy can further improve low-resource NMT with data from other auxiliary languages. In this paper, we seek to construct a sampling distribution over all multilingual data, so that it minimizes the training loss of the low-resource language. Based on this formulation, we propose an efficient algorithm, Target Conditioned Sampling (TCS), which first samples a target sentence, and then conditionally samples its source sentence. Experiments show that TCS brings significant gains of up to 2 BLEU on three of four languages we test, with minimal training overhead." @default.
- W2945501029 created "2019-05-29" @default.
- W2945501029 creator A5040523461 @default.
- W2945501029 creator A5068811427 @default.
- W2945501029 date "2019-05-20" @default.
- W2945501029 modified "2023-09-27" @default.
- W2945501029 title "Target Conditioned Sampling: Optimizing Data Selection for Multilingual Neural Machine Translation" @default.
- W2945501029 cites W1905522558 @default.
- W2945501029 cites W2117278770 @default.
- W2945501029 cites W2130942839 @default.
- W2945501029 cites W2144600658 @default.
- W2945501029 cites W2147262247 @default.
- W2945501029 cites W22168010 @default.
- W2945501029 cites W2740718109 @default.
- W2945501029 cites W2740743644 @default.
- W2945501029 cites W2752630748 @default.
- W2945501029 cites W2756566411 @default.
- W2945501029 cites W2887920589 @default.
- W2945501029 cites W2888519496 @default.
- W2945501029 cites W2962784628 @default.
- W2945501029 cites W2963088995 @default.
- W2945501029 cites W2963247703 @default.
- W2945501029 cites W2963250244 @default.
- W2945501029 cites W2963366389 @default.
- W2945501029 cites W2963831310 @default.
- W2945501029 cites W2963993537 @default.
- W2945501029 cites W2964085268 @default.
- W2945501029 cites W630532510 @default.
- W2945501029 doi "https://doi.org/10.48550/arxiv.1905.08212" @default.
- W2945501029 hasPublicationYear "2019" @default.
- W2945501029 type Work @default.
- W2945501029 sameAs 2945501029 @default.
- W2945501029 citedByCount "2" @default.
- W2945501029 countsByYear W29455010292020 @default.
- W2945501029 countsByYear W29455010292021 @default.
- W2945501029 crossrefType "posted-content" @default.
- W2945501029 hasAuthorship W2945501029A5040523461 @default.
- W2945501029 hasAuthorship W2945501029A5068811427 @default.
- W2945501029 hasBestOaLocation W29455010291 @default.
- W2945501029 hasConcept C104317684 @default.
- W2945501029 hasConcept C105580179 @default.
- W2945501029 hasConcept C106131492 @default.
- W2945501029 hasConcept C119857082 @default.
- W2945501029 hasConcept C140779682 @default.
- W2945501029 hasConcept C149364088 @default.
- W2945501029 hasConcept C154945302 @default.
- W2945501029 hasConcept C185592680 @default.
- W2945501029 hasConcept C199360897 @default.
- W2945501029 hasConcept C203005215 @default.
- W2945501029 hasConcept C204321447 @default.
- W2945501029 hasConcept C206345919 @default.
- W2945501029 hasConcept C2777530160 @default.
- W2945501029 hasConcept C2779960059 @default.
- W2945501029 hasConcept C2780801425 @default.
- W2945501029 hasConcept C31258907 @default.
- W2945501029 hasConcept C31972630 @default.
- W2945501029 hasConcept C41008148 @default.
- W2945501029 hasConcept C55493867 @default.
- W2945501029 hasConcept C81917197 @default.
- W2945501029 hasConceptScore W2945501029C104317684 @default.
- W2945501029 hasConceptScore W2945501029C105580179 @default.
- W2945501029 hasConceptScore W2945501029C106131492 @default.
- W2945501029 hasConceptScore W2945501029C119857082 @default.
- W2945501029 hasConceptScore W2945501029C140779682 @default.
- W2945501029 hasConceptScore W2945501029C149364088 @default.
- W2945501029 hasConceptScore W2945501029C154945302 @default.
- W2945501029 hasConceptScore W2945501029C185592680 @default.
- W2945501029 hasConceptScore W2945501029C199360897 @default.
- W2945501029 hasConceptScore W2945501029C203005215 @default.
- W2945501029 hasConceptScore W2945501029C204321447 @default.
- W2945501029 hasConceptScore W2945501029C206345919 @default.
- W2945501029 hasConceptScore W2945501029C2777530160 @default.
- W2945501029 hasConceptScore W2945501029C2779960059 @default.
- W2945501029 hasConceptScore W2945501029C2780801425 @default.
- W2945501029 hasConceptScore W2945501029C31258907 @default.
- W2945501029 hasConceptScore W2945501029C31972630 @default.
- W2945501029 hasConceptScore W2945501029C41008148 @default.
- W2945501029 hasConceptScore W2945501029C55493867 @default.
- W2945501029 hasConceptScore W2945501029C81917197 @default.
- W2945501029 hasLocation W29455010291 @default.
- W2945501029 hasOpenAccess W2945501029 @default.
- W2945501029 hasPrimaryLocation W29455010291 @default.
- W2945501029 hasRelatedWork W170118632 @default.
- W2945501029 hasRelatedWork W1947252915 @default.
- W2945501029 hasRelatedWork W1964392734 @default.
- W2945501029 hasRelatedWork W1978971213 @default.
- W2945501029 hasRelatedWork W2359582629 @default.
- W2945501029 hasRelatedWork W2398825887 @default.
- W2945501029 hasRelatedWork W2507994462 @default.
- W2945501029 hasRelatedWork W2606032440 @default.
- W2945501029 hasRelatedWork W2946260846 @default.
- W2945501029 hasRelatedWork W4287208387 @default.
- W2945501029 isParatext "false" @default.
- W2945501029 isRetracted "false" @default.
- W2945501029 magId "2945501029" @default.
- W2945501029 workType "article" @default.