Proficiency Matters Quality Estimation in Grammatical Error Correction

TL;DR

This paper examines how the proficiency level of language learners impacts the effectiveness of quality estimation models in grammatical error correction, emphasizing the need for proficiency-aware evaluation for robustness.

Contribution

It introduces a new QE dataset with multiple proficiency levels and demonstrates the importance of proficiency-wise evaluation for more reliable GEC quality estimation.

Findings

Proficiency levels influence QE model performance.

Proficiency-wise evaluation improves model robustness.

Existing models are biased toward high-proficiency data.

Abstract

This study investigates how supervised quality estimation (QE) models of grammatical error correction (GEC) are affected by the learners' proficiency with the data. QE models for GEC evaluations in prior work have obtained a high correlation with manual evaluations. However, when functioning in a real-world context, the data used for the reported results have limitations because prior works were biased toward data by learners with relatively high proficiency levels. To address this issue, we created a QE dataset that includes multiple proficiency levels and explored the necessity of performing proficiency-wise evaluation for QE of GEC. Our experiments demonstrated that differences in evaluation dataset proficiency affect the performance of QE models, and proficiency-wise evaluation helps create more robust models.