Enhancing Automated Essay Scoring with Three Techniques: Two-Stage Fine-Tuning, Score Alignment, and Self-Training
Hongseok Choi, Serynn Kim, Wencke Liermann, Jin Seong, Jin-Xia Huang · 2026
Read Enhancing Automated Essay Scoring with Three Techniques: Two-Stage Fine-Tuning, Score Alignment, and Self-Training by Hongseok Choi, Serynn Kim, Wencke Liermann, Jin Seong, Jin-Xia Huang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01747.txt.
Description
Hongseok Choi and colleagues improve essay scoring, where scarcity of labeled data blocks robust systems in practice. They add three techniques to DualBERT: two stage fine tuning with low rank adaptations for essays. Score alignment matches score distributions, and uncertainty aware self training on unlabeled data adds pseudo labeled samples while limiting noise spread. Tests on ASAP++, TOEFL11 and ELLIPSE show all three techniques help with thirty two samples. Their integration reaches over ninety percent of full data performance, and score alignment helps in both limited and full data settings.
Catalog record
| CREATOR | Hongseok Choi, Serynn Kim, Wencke Liermann, Jin Seong, Jin-Xia Huang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 42872 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.01747