| 模型 | NLPCC2016 CKBQA | CCKS 2019EL | ||||
| P | R | F1 | P | R | F1 | |
| State-of-the-Art | ||||||
| BB-KBQA [7] | 96.96 | 97.02 | 96.99 | - | - | - |
| 方法 [8] | - | - | - | 82.68 | 85.34 | 83.98 |
| 知识增强的预训练模型 | ||||||
| K-BERT [19] | 97.21 | 97.6 | 97.4 | 81.26 | 84.32 | 82.76 |
| EKBERT | ||||||
| EKBERT-MD | 98.87 | 98.87 | 98.87 | 86.09 | 88.07 | 87.06 |