Towards Prompt Generalization: Grammar-aware Cross-Prompt Automated Essay Scoring

TL;DR

This paper introduces GAPS, a grammar-aware method for automated essay scoring that improves generalization across unseen prompts by focusing on prompt-independent syntactic features and grammatical correctness.

Contribution

The paper proposes a novel grammar-aware approach that captures prompt-independent features for cross-prompt essay scoring, enhancing generalization and robustness.

Findings

GAPS significantly improves prompt-independent trait scoring.

The method achieves higher QWK scores in cross-prompt scenarios.

Empirical results validate the effectiveness of grammar-aware features.

Abstract

In automated essay scoring (AES), recent efforts have shifted toward cross-prompt settings that score essays on unseen prompts for practical applicability. However, prior methods trained with essay-score pairs of specific prompts pose challenges in obtaining prompt-generalized essay representation. In this work, we propose a grammar-aware cross-prompt trait scoring (GAPS), which internally captures prompt-independent syntactic aspects to learn generic essay representation. We acquire grammatical error-corrected information in essays via the grammar error correction technique and design the AES model to seamlessly integrate such information. By internally referring to both the corrected and the original essays, the model can focus on generic features during training. Empirical experiments validate our method's generalizability, showing remarkable improvements in prompt-independent and grammar-related traits. Furthermore, GAPS achieves notable QWK gains in the most challenging cross-prompt scenario, highlighting its strength in evaluating unseen prompts.