arXiv:2601.16724 · v3

Mitigating Bias in Automated Essay Scoring for ESL Learners via Contrastive Learning

首次提交 2026-01-23 21:18

版本更新 2026-11-02 05:12

雷达观测 2026-11-04 23:42

摘要

Automated Essay Scoring systems disproportionately penalize high-proficiency English as a Second Language (ESL) learners. We propose Contrastive Learning with Matched Essay Pairs (CL-MEP), a bi-directional alignment strategy. CL-MEP reduces this scoring bias by 39.9% while improving overall accuracy, successfully disentangling valid syntactic complexity from surface-level grammatical errors.