FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3098989509> ?p ?o ?g. }

• W3098989509 abstract "Exposing diverse subword segmentations to neural machine translation (NMT) models often improves the robustness of machine translation as NMT models can experience various subword candidates. However, the diversification of subword segmentations mostly relies on the pre-trained subword language models from which erroneous segmentations of unseen words are less likely to be sampled. In this paper, we present adversarial subword regularization (ADVSR) to study whether gradient signals during training can be a substitute criterion for exposing diverse subword segmentations. We experimentally show that our model-based adversarial samples effectively encourage NMT models to be less sensitive to segmentation errors and improve the performance of NMT models in low-resource and out-domain datasets." @default.
• W3098989509 created "2020-11-23" @default.
• W3098989509 creator A5002413587 @default.
• W3098989509 creator A5020746366 @default.
• W3098989509 creator A5075876055 @default.
• W3098989509 creator A5076917278 @default.
• W3098989509 date "2020-01-01" @default.
• W3098989509 modified "2023-09-23" @default.
• W3098989509 title "Adversarial Subword Regularization for Robust Neural Machine Translation" @default.
• W3098989509 cites W1991133427 @default.
• W3098989509 cites W2008225289 @default.
• W3098989509 cites W2049633694 @default.
• W3098989509 cites W2101105183 @default.
• W3098989509 cites W2121879602 @default.
• W3098989509 cites W2157381218 @default.
• W3098989509 cites W222053410 @default.
• W3098989509 cites W2325237720 @default.
• W3098989509 cites W2531207078 @default.
• W3098989509 cites W2735135478 @default.
• W3098989509 cites W2799194071 @default.
• W3098989509 cites W2896691342 @default.
• W3098989509 cites W2922293812 @default.
• W3098989509 cites W2933138175 @default.
• W3098989509 cites W2938348542 @default.
• W3098989509 cites W2946232455 @default.
• W3098989509 cites W2950651087 @default.
• W3098989509 cites W2962784628 @default.
• W3098989509 cites W2963207607 @default.
• W3098989509 cites W2963250244 @default.
• W3098989509 cites W2963324947 @default.
• W3098989509 cites W2963341956 @default.
• W3098989509 cites W2963403868 @default.
• W3098989509 cites W2963532001 @default.
• W3098989509 cites W2963661177 @default.
• W3098989509 cites W2963823140 @default.
• W3098989509 cites W2963831310 @default.
• W3098989509 cites W2963887123 @default.
• W3098989509 cites W2963969878 @default.
• W3098989509 cites W2963979492 @default.
• W3098989509 cites W2964247056 @default.
• W3098989509 cites W2964265128 @default.
• W3098989509 cites W2970510999 @default.
• W3098989509 cites W2970558573 @default.
• W3098989509 cites W2970597249 @default.
• W3098989509 cites W2971134989 @default.
• W3098989509 cites W2982756474 @default.
• W3098989509 cites W2984051011 @default.
• W3098989509 cites W3001816066 @default.
• W3098989509 cites W3035207248 @default.
• W3098989509 cites W3105718208 @default.
• W3098989509 doi "https://doi.org/10.18653/v1/2020.findings-emnlp.175" @default.
• W3098989509 hasPublicationYear "2020" @default.
• W3098989509 type Work @default.
• W3098989509 sameAs 3098989509 @default.
• W3098989509 citedByCount "0" @default.
• W3098989509 crossrefType "proceedings-article" @default.
• W3098989509 hasAuthorship W3098989509A5002413587 @default.
• W3098989509 hasAuthorship W3098989509A5020746366 @default.
• W3098989509 hasAuthorship W3098989509A5075876055 @default.
• W3098989509 hasAuthorship W3098989509A5076917278 @default.
• W3098989509 hasBestOaLocation W30989895091 @default.
• W3098989509 hasConcept C104317684 @default.
• W3098989509 hasConcept C105580179 @default.
• W3098989509 hasConcept C119857082 @default.
• W3098989509 hasConcept C137293760 @default.
• W3098989509 hasConcept C149364088 @default.
• W3098989509 hasConcept C154945302 @default.
• W3098989509 hasConcept C185592680 @default.
• W3098989509 hasConcept C203005215 @default.
• W3098989509 hasConcept C204321447 @default.
• W3098989509 hasConcept C2776135515 @default.
• W3098989509 hasConcept C28490314 @default.
• W3098989509 hasConcept C2984842247 @default.
• W3098989509 hasConcept C37736160 @default.
• W3098989509 hasConcept C41008148 @default.
• W3098989509 hasConcept C50644808 @default.
• W3098989509 hasConcept C55493867 @default.
• W3098989509 hasConcept C63479239 @default.
• W3098989509 hasConcept C89600930 @default.
• W3098989509 hasConceptScore W3098989509C104317684 @default.
• W3098989509 hasConceptScore W3098989509C105580179 @default.
• W3098989509 hasConceptScore W3098989509C119857082 @default.
• W3098989509 hasConceptScore W3098989509C137293760 @default.
• W3098989509 hasConceptScore W3098989509C149364088 @default.
• W3098989509 hasConceptScore W3098989509C154945302 @default.
• W3098989509 hasConceptScore W3098989509C185592680 @default.
• W3098989509 hasConceptScore W3098989509C203005215 @default.
• W3098989509 hasConceptScore W3098989509C204321447 @default.
• W3098989509 hasConceptScore W3098989509C2776135515 @default.
• W3098989509 hasConceptScore W3098989509C28490314 @default.
• W3098989509 hasConceptScore W3098989509C2984842247 @default.
• W3098989509 hasConceptScore W3098989509C37736160 @default.
• W3098989509 hasConceptScore W3098989509C41008148 @default.
• W3098989509 hasConceptScore W3098989509C50644808 @default.
• W3098989509 hasConceptScore W3098989509C55493867 @default.
• W3098989509 hasConceptScore W3098989509C63479239 @default.
• W3098989509 hasConceptScore W3098989509C89600930 @default.
• W3098989509 hasLocation W30989895091 @default.
• W3098989509 hasLocation W30989895092 @default.
• W3098989509 hasOpenAccess W3098989509 @default.

• next