Knowledge Graph Completion on FB60K-NYT10 50% training ratio (test)
48.98Hits@5TagReal
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TagRealModel Category=PLM-based2023.05 | 48.98 | 55.64 | 38.03 | |
| PKGCModel Category=PLM-based2023.05 | 41.93 | 46.7 | 31.81 | |
| TuckERModel Category=KGE-based2023.05 | 41.73 | 45.58 | 33.84 | |
| TransEModel Category=KGE-based2023.05 | 41.54 | 45.74 | 25.82 | |
| ConvEModel Category=KGE-based2023.05 | 40.1 | 44.03 | 32.97 | |
| RotatEModel Category=KGE-based2023.05 | 35.48 | 39.42 | 28.92 | |
| JointNREModel Category=Text&KGE-based2023.05 | 26.96 | 31.54 | 21.24 | |
| CPLModel Category=RL-based2023.05 | 26.81 | 31.7 | 23.8 | |
| MINERVAModel Category=RL-based2023.05 | 25.16 | 31.54 | 22.24 | |
| TransE+LineModel Category=Text&KGE-based2023.05 | 21.7 | 25.75 | 8.81 | |
| DisMultModel Category=KGE-based2023.05 | 15.98 | 18.85 | 13.14 | |
| ComplExModel Category=KGE-based2023.05 | 15 | 17.73 | 12.21 | |
| RC-NetModel Category=Text&KGE-based2023.05 | 14.87 | 16.54 | 14.63 |