Refining Corpora from a Model Calibration Perspective for Chinese Spelling Correction
This work addresses data quality issues in CSC, a domain-specific task, by refining existing augmentation methods to improve model calibration and reduce over-correction.
The paper tackled the problem of noisy data in Chinese Spelling Correction (CSC) corpora, which can cause over-correction, by proposing a corpus refining strategy that filters OCR/ASR-based data using a well-calibrated model, resulting in state-of-the-art performance on three benchmarks and reduced false positives.
Chinese Spelling Correction (CSC) commonly lacks large-scale high-quality corpora, due to the labor-intensive labeling of spelling errors in real-life human writing or typing scenarios. Two data augmentation methods are widely adopted: (1) \textit{Random Replacement} with the guidance of confusion sets and (2) \textit{OCR/ASR-based Generation} that simulates character misusing. However, both methods inevitably introduce noisy data (e.g., false spelling errors), potentially leading to over-correction. By carefully analyzing the two types of corpora, we find that though the latter achieves more robust generalization performance, the former yields better-calibrated CSC models. We then provide a theoretical analysis of this empirical observation, based on which a corpus refining strategy is proposed. Specifically, OCR/ASR-based data samples are fed into a well-calibrated CSC model trained on random replacement-based corpora and then filtered based on prediction confidence. By learning a simple BERT-based model on the refined OCR/ASR-based corpus, we set up impressive state-of-the-art performance on three widely-used benchmarks, while significantly alleviating over-correction (e.g., lowering false positive predictions).