Dialogue-based Multiple-choice Question Answering on DREAM (test)
91.8AccuracyALBERT_xxlarge + DUMA
Evaluation results
| Method | Links | |
|---|---|---|
| ALBERT_xxlarge + DUMABackbone=ALBERT-xxlarge, training_mode=multi-task learning2020.01 | 91.8 | |
| ALBERT_xxlarge + DUMABackbone=ALBERT-xxlarge2020.01 | 90.5 | |
| RoBERTa_large + MMMBackbone=RoBERTa-large, Augmentation=MMM2020.01 | 88.9 | |
| ALBERT_xxlargeBackbone=ALBERT-xxlarge2020.01 | 88.5 | |
| PathReasonerBackbone=RoBERTa-Large2024.05 | 86.84 | |
| RoBERTa-LBackbone=RoBERTa-Large2024.05 | 84.74 | |
| LogiformerBackbone=RoBERTa-Large2024.05 | 83.76 | |
| ALIGNModel Variant=large, Parameters=355M, Evaluation Protocol=Zero-shot2023.07 | 81.1 | |
| FLAN-T5Model Variant=large, Parameters=780M, Evaluation Protocol=Zero-shot2023.07 | 79 | |
| XLNetsource=leaderboard2019.10 | 72 | |
| XLNet_largeBackbone=XLNet-large, source=leaderboard2020.01 | 72 | |
| ALIGNModel Variant=base, Parameters=125M, Evaluation Protocol=Zero-shot2023.07 | 71.3 | |
| OCN + AT injectionknowledge_source=ATOMIC2019.10 | 70.1 | |
| FLAN-T5Model Variant=base, Parameters=250M, Evaluation Protocol=Zero-shot2023.07 | 69.9 | |
| OCN2019.10 | 69.8 | |
| OCN + CN injectionknowledge_source=ConceptNet2019.10 | 69.6 | |
| BERT_large + WAEBackbone=BERT-large, Augmentation=WAE, source=leaderboard2020.01 | 69 | |
| BERT Largesource=leaderboard2019.10 | 66.8 | |
| BERT_largeBackbone=BERT-large2020.01 | 66.8 | |
| OCN + OMCS pre-trainpre-training=OMCS2019.10 | 62.6 | |
| OCN + ATOMIC pre-trainpre-training=ATOMIC2019.10 | 58.8 |