Reaching Human-level Performance in Automatic Grammatical Error Correction: An Empirical Study
This solves the problem of automated grammatical error correction for language learners and writers, representing a significant advance rather than an incremental improvement.
The paper tackles grammatical error correction by proposing a fluency boost learning and inference mechanism with convolutional seq2seq models, achieving state-of-the-art performance of 75.72 F0.5 on CoNLL-2014 and 62.42 GLEU on JFLEG, reaching human-level benchmarks.
Neural sequence-to-sequence (seq2seq) approaches have proven to be successful in grammatical error correction (GEC). Based on the seq2seq framework, we propose a novel fluency boost learning and inference mechanism. Fluency boosting learning generates diverse error-corrected sentence pairs during training, enabling the error correction model to learn how to improve a sentence's fluency from more instances, while fluency boosting inference allows the model to correct a sentence incrementally with multiple inference steps. Combining fluency boost learning and inference with convolutional seq2seq models, our approach achieves the state-of-the-art performance: 75.72 (F_{0.5}) on CoNLL-2014 10 annotation dataset and 62.42 (GLEU) on JFLEG test set respectively, becoming the first GEC system that reaches human-level performance (72.58 for CoNLL and 62.37 for JFLEG) on both of the benchmarks.