Automated Essay Scoring on ASAP++ 32-data setting (test)
0.635QWK (P1)DualBERT + LoRA + SA + UST
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| DualBERT + LoRA + SA + USTtraining_setting=32-data, component=Integrated Method2026.02 | 0.635 | 0.661 | 0.631 | 0.685 | 0.667 | 0.67 | 0.7 | 0.571 | 0.653 | 0.011 | |
| DualBERT + USTtraining_setting=32-data, component=Uncertainty-aware Self-Training (UST)2026.02 | 0.625 | 0.613 | 0.557 | 0.653 | 0.619 | 0.608 | 0.665 | 0.576 | 0.615 | 0.008 | |
| DualBERT + SAtraining_setting=32-data, component=Score Alignment (SA)2026.02 | 0.623 | 0.64 | 0.602 | 0.665 | 0.648 | 0.645 | 0.684 | 0.592 | 0.638 | 0.014 | |
| DualBERT + LoRAtraining_setting=32-data, component=LoRA (Two-Stage fine-tuning)2026.02 | 0.621 | 0.617 | 0.544 | 0.651 | 0.605 | 0.577 | 0.661 | 0.576 | 0.607 | 0.017 | |
| DualBERTtraining_setting=32-data, base_model=DualBERT2026.02 | 0.614 | 0.611 | 0.528 | 0.629 | 0.599 | 0.578 | 0.653 | 0.582 | 0.599 | 0.018 | |
| ArTS+RMTStraining_setting=32-data, reimplemented=true2026.02 | 0.479 | 0.416 | 0.494 | 0.562 | 0.543 | 0.475 | 0.374 | 0.349 | 0.462 | 0.076 |