模型 | NLPCC 2016CKBQA | CCKS 2019EL | |||||
Top 1 | Top 2 | Top 3 | MRR | Top 1 | Top 2 | Top 3 | |
State-of-the-Art | |||||||
BB-KBQA [7] | 89.14 | 93.19 | 95.05 | 92.16 | - | - | - |
方法 [8] | - | - | - | - | 93.93 | 99.09 | 99.86 |
知识增强的预训练模型 | |||||||
K-BERT [19] | 89.58 | 92.58 | 94.41 | 91.46 | 93.38 | 98.16 | 99.21 |
EKBERT | |||||||
EKBERT-ED-max | 93.39 | 96.98 | 98.4 | 95.94 | 93.83 | 99.14 | 99.82 |
EKBERT-ED-ave | 93.28 | 96.79 | 98.36 | 95.87 | 93.76 | 99.02 | 99.79 |