Neural Automated Writing Evaluation with Corrective Feedback

TL;DR

The paper addresses the need for efficient, immediate feedback in second language writing by integrating neural automated writing evaluation (AWE) with grammatical error correction (GEC). It combines a denoised ASAP/ASAP++ based AWE component with a seq2seq/ensemble GEC module trained on BEA2019 data to produce inline corrections and multi-rubric scores, enabling simulated examination practice. The system supports teacher involvement, cross-language expansion through multilingual rubrics, and planned perception studies to assess usability and fairness. It achieves competitive GEC performance (best BERT-based model with $F_{0.5}=65.29$ on BEA2019) and demonstrates a concrete path toward scalable, language-agnostic evaluation with corrective feedback that reduces instructor workload.

Abstract

The utilization of technology in second language learning and teaching has become ubiquitous. For the assessment of writing specifically, automated writing evaluation (AWE) and grammatical error correction (GEC) have become immensely popular and effective methods for enhancing writing proficiency and delivering instant and individualized feedback to learners. By leveraging the power of natural language processing (NLP) and machine learning algorithms, AWE and GEC systems have been developed separately to provide language learners with automated corrective feedback and more accurate and unbiased scoring that would otherwise be subject to examiners. In this paper, we propose an integrated system for automated writing evaluation with corrective feedback as a means of bridging the gap between AWE and GEC results for second language learners. This system enables language learners to simulate the essay writing tests: a student writes and submits an essay, and the system returns the assessment of the writing along with suggested grammatical error corrections. Given that automated scoring and grammatical correction are more efficient and cost-effective than human grading, this integrated system would also alleviate the burden of manually correcting innumerable essays.

Figures (3)

• Figure 1: System workflow of integrated AWE and GEC for language learners. A user flow can be applied to simulate examination situations: the language learners receive instant objective scoring results and corrective feedback.
• Figure 2: Interface screenshot for AWE and GEC results: tokens in the faded red rectangles have been deleted; tokens in the green rectangles are corresponding grammatical corrections inserted by GEC; AWE scores are scaled to 0-100.
• Figure 3: Future system workflow for the language learners and the instructors