A Self-Refinement Strategy for Noise Reduction in Grammatical Error Correction
This addresses data quality issues in grammatical error correction, which is important for NLP applications like writing assistance, but is incremental as it builds on existing denoising methods.
The paper tackles the problem of noise in grammatical error correction datasets by proposing a self-refinement method to denoise them, achieving state-of-the-art performance on benchmarks like CoNLL-2014, JFLEG, and BEA-2019 with improved recall and overall metrics.
Existing approaches for grammatical error correction (GEC) largely rely on supervised learning with manually created GEC datasets. However, there has been little focus on verifying and ensuring the quality of the datasets, and on how lower-quality data might affect GEC performance. We indeed found that there is a non-negligible amount of "noise" where errors were inappropriately edited or left uncorrected. To address this, we designed a self-refinement method where the key idea is to denoise these datasets by leveraging the prediction consistency of existing models, and outperformed strong denoising baseline methods. We further applied task-specific techniques and achieved state-of-the-art performance on the CoNLL-2014, JFLEG, and BEA-2019 benchmarks. We then analyzed the effect of the proposed denoising method, and found that our approach leads to improved coverage of corrections and facilitated fluency edits which are reflected in higher recall and overall performance.