Erroneous data generation for Grammatical Error Correction
Shuyao Xu, Jiehao Zhang, Jin Chen, Long Qin · 2019
It has been demonstrated that the utilization of a monolingual corpus in neural Grammatical Error Correction (GEC) systems can significantly improve the system performance.The previous state-of-theart neural GEC system is an ensemble of four Transformer models pretrained on a large amount of Wikipedia Edits.The Singsound GEC system follows a similar approach but is equipped with a sophisticated erroneous data generating component.Our system achieved an F 0.5 of 66.61 in the BEA 2019 Shared Task: Grammatical Error Correction.With our novel erroneous data generating component, the Singsound neural GEC system yielded an M 2 of 63.2 on the CoNLL-2014 benchmark (8.4% relative improvement over the previous state-of-the-art system).