Machine Reading Comprehension on MultiRC SuperGLUE (test)
27.2EMKELM Both
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| KELM BothKnowledge Source=WordNet + NELL2021.09 | 27.2 | 70.8 | |
| KELM NELLKnowledge Source=NELL2021.09 | 26.5 | 70.6 | |
| KELM WordNetKnowledge Source=WordNet2021.09 | 25.9 | 69.2 | |
| KT-NET BOTHKnowledge Source=WordNet + NELL, Implementation=Authors' own implementation2021.09 | 25.4 | 71.1 | |
| BERTlargeBackbone=BERT-Large2021.09 | 24.1 | 70 |