FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2948335087> ?p ?o ?g. }

• W2948335087 abstract "Considerable effort has been made to address the data sparsity problem in neural grammatical error correction. In this work, we propose a simple and surprisingly effective unsupervised synthetic error generation method based on confusion sets extracted from a spellchecker to increase the amount of training data. Synthetic data is used to pre-train a Transformer sequence-to-sequence model, which not only improves over a strong baseline trained on authentic error-annotated data, but also enables the development of a practical GEC system in a scenario where little genuine error-annotated data is available. The developed systems placed first in the BEA19 shared task, achieving 69.47 and 64.24 F0.5 in the restricted and low-resource tracks respectively, both on the W&I+LOCNESS test set. On the popular CoNLL 2014 test set, we report state-of-the-art results of 64.16 M² for the submitted system, and 61.30 M² for the constrained system trained on the NUCLE and Lang-8 data." @default.
• W2948335087 created "2019-06-14" @default.
• W2948335087 creator A5023392498 @default.
• W2948335087 creator A5065193578 @default.
• W2948335087 creator A5070589644 @default.
• W2948335087 date "2019-01-01" @default.
• W2948335087 modified "2023-10-14" @default.
• W2948335087 title "Neural Grammatical Error Correction Systems with Unsupervised Pre-training on Synthetic Data" @default.
• W2948335087 cites W1520449809 @default.
• W2948335087 cites W1647671624 @default.
• W2948335087 cites W1665214252 @default.
• W2948335087 cites W1721115786 @default.
• W2948335087 cites W2033922434 @default.
• W2948335087 cites W2100495367 @default.
• W2948335087 cites W2117278770 @default.
• W2948335087 cites W2124725212 @default.
• W2948335087 cites W2125257468 @default.
• W2948335087 cites W2132782512 @default.
• W2948335087 cites W2134800885 @default.
• W2948335087 cites W2140372282 @default.
• W2948335087 cites W2144555877 @default.
• W2948335087 cites W2146574666 @default.
• W2948335087 cites W2153013403 @default.
• W2948335087 cites W2153579005 @default.
• W2948335087 cites W2170527467 @default.
• W2948335087 cites W2252000648 @default.
• W2948335087 cites W2315316408 @default.
• W2948335087 cites W2538523173 @default.
• W2948335087 cites W2555428947 @default.
• W2948335087 cites W2589277916 @default.
• W2948335087 cites W2740433069 @default.
• W2948335087 cites W2741494657 @default.
• W2948335087 cites W2759575900 @default.
• W2948335087 cites W2798416860 @default.
• W2948335087 cites W2803237843 @default.
• W2948335087 cites W2805583812 @default.
• W2948335087 cites W2810035278 @default.
• W2948335087 cites W2888555799 @default.
• W2948335087 cites W2899310090 @default.
• W2948335087 cites W2903193068 @default.
• W2948335087 cites W2925188774 @default.
• W2948335087 cites W2936597270 @default.
• W2948335087 cites W2962801832 @default.
• W2948335087 cites W2962863357 @default.
• W2948335087 cites W2963109131 @default.
• W2948335087 cites W2963250244 @default.
• W2948335087 cites W2963266340 @default.
• W2948335087 cites W2963267786 @default.
• W2948335087 cites W2963347649 @default.
• W2948335087 cites W2963403868 @default.
• W2948335087 cites W2963881719 @default.
• W2948335087 cites W2963975242 @default.
• W2948335087 cites W2963979492 @default.
• W2948335087 cites W2964082031 @default.
• W2948335087 cites W2964099746 @default.
• W2948335087 cites W2964121744 @default.
• W2948335087 cites W2964187553 @default.
• W2948335087 cites W2964258094 @default.
• W2948335087 cites W2970868759 @default.
• W2948335087 cites W43022990 @default.
• W2948335087 doi "https://doi.org/10.18653/v1/w19-4427" @default.
• W2948335087 hasPublicationYear "2019" @default.
• W2948335087 type Work @default.
• W2948335087 sameAs 2948335087 @default.
• W2948335087 citedByCount "118" @default.
• W2948335087 countsByYear W29483350872019 @default.
• W2948335087 countsByYear W29483350872020 @default.
• W2948335087 countsByYear W29483350872021 @default.
• W2948335087 countsByYear W29483350872022 @default.
• W2948335087 countsByYear W29483350872023 @default.
• W2948335087 crossrefType "proceedings-article" @default.
• W2948335087 hasAuthorship W2948335087A5023392498 @default.
• W2948335087 hasAuthorship W2948335087A5065193578 @default.
• W2948335087 hasAuthorship W2948335087A5070589644 @default.
• W2948335087 hasBestOaLocation W29483350871 @default.
• W2948335087 hasConcept C103088060 @default.
• W2948335087 hasConcept C11171543 @default.
• W2948335087 hasConcept C11413529 @default.
• W2948335087 hasConcept C119857082 @default.
• W2948335087 hasConcept C121332964 @default.
• W2948335087 hasConcept C153180895 @default.
• W2948335087 hasConcept C154945302 @default.
• W2948335087 hasConcept C15744967 @default.
• W2948335087 hasConcept C160920958 @default.
• W2948335087 hasConcept C162324750 @default.
• W2948335087 hasConcept C165801399 @default.
• W2948335087 hasConcept C16910744 @default.
• W2948335087 hasConcept C169903167 @default.
• W2948335087 hasConcept C177264268 @default.
• W2948335087 hasConcept C187736073 @default.
• W2948335087 hasConcept C199360897 @default.
• W2948335087 hasConcept C204321447 @default.
• W2948335087 hasConcept C2780451532 @default.
• W2948335087 hasConcept C2781140086 @default.
• W2948335087 hasConcept C28490314 @default.
• W2948335087 hasConcept C41008148 @default.
• W2948335087 hasConcept C51632099 @default.
• W2948335087 hasConcept C62520636 @default.
• W2948335087 hasConcept C66322947 @default.
• W2948335087 hasConceptScore W2948335087C103088060 @default.

• next