Towards Explainable Chinese Native Learner Essay Fluency Assessment: Dataset, Tasks, and Method

Anonymous

Abstract: Grammatical Error Correction (GEC) is a crucial technique in Automated Essay Assessment (AEA) for evaluating the fluency of essays. However, in Chinese, existing GEC datasets often fail to consider the importance of specific grammatical error types within compositional scenarios, lack research on data collected from native Chinese speakers, and largely overlook cross-sentence grammatical errors. To address these issues, we present CEFGEC (Chinese Essay Fluency Grammatical Error Correction), an extensive corpus that focuses on fine-grained and multi-dimensional fluency analysis. Furthermore, we propose a novel Grammatical Error Identification and Correction via Knowledge Distillation (GEIC-KD) model to investigate the relationships between multi-dimensional annotated content. Compared to other benchmark models, experimental results illustrate that GEIC-KD outperforms them on our dataset. Our findings also further emphasize the importance of fine-grained annotations in fluency assessment. We will make the corpus and related codes available for research.

Paper Type: long

Research Area: NLP Applications

Contribution Types: Model analysis & interpretability, Data resources, Data analysis

Languages Studied: Chinese