Machine Reading Comprehension on DREAM (test)
91.8AccuracyALBERTxxlarge + DUMA + MTL
Evaluation results
| Method | Links | |
|---|---|---|
| ALBERTxxlarge + DUMA + MTLBackbone=ALBERT-xxlarge, Module=DUMA, Training Strategy=Multi-task learning (MTL)2022.03 | 91.8 | |
| ALBERTxxlarge + DUMABackbone=ALBERT-xxlarge2022.03 | 90.4 | |
| ALBERTxxlarge + DUMABackbone=ALBERT-xxlarge, Module=DUMA2022.03 | 90.4 | |
| POI-NetBackbone=ALBERT-xxlarge2022.03 | 90.3 | |
| POI-Net on ALBERTxxlargeBackbone=ALBERT-xxlarge2022.03 | 90.3 | |
| RoBERTalarge+MMMBackbone=RoBERTa-large2022.03 | 88.9 | |
| RoBERTalarge + MMMBackbone=RoBERTa-large, External Resource=MMM2022.03 | 88.9 | |
| ALBERTxxlargeBackbone=ALBERT-xxlarge2022.03 | 88.5 | |
| ALBERTxxlargeBackbone=ALBERT-xxlarge, rerun=true2022.03 | 88.5 | |
| ALBERTxxlargeBackbone=ALBERT-xxlarge, rerun=true2022.03 | 88.3 | |
| RoBERTalargeBackbone=RoBERTa-large2022.03 | 85 | |
| RoBERTalargeBackbone=RoBERTa-large2022.03 | 85 | |
| XLNetlargeBackbone=XLNet-large2022.03 | 72 | |
| POI-NetBackbone=ALBERT-base2022.03 | 68.5 | |
| POI-Net on ALBERTbaseBackbone=ALBERT-base2022.03 | 68.5 | |
| BERTlargeBackbone=BERT-large2022.03 | 66.8 | |
| BERTlargeBackbone=BERT-large2022.03 | 66.8 | |
| ALBERTbaseBackbone=ALBERT-base, rerun=true2022.03 | 65.6 | |
| ALBERTbaseBackbone=ALBERT-base, rerun=true2022.03 | 65.6 | |
| ALBERTbaseBackbone=ALBERT-base2022.03 | 64.4 | |
| BERTbaseBackbone=BERT-base2022.03 | 63.2 | |
| BERTbaseBackbone=BERT-base2022.03 | 63.2 | |
| FTLM++2022.03 | 58.2 |