Knowledge Graph Reasoning on FB15k-237 (test)
23.4HITS@3 (Avg)ConE
Evaluation results
| Method | Links | ||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ConEembedding dimension (d)=8002021.10 | 23.4 | — | — | — | 32.6 | 47.3 | 14.5 | — | — | — | 41.8 | 12.8 | 11 | 25.5 | 14 | 10.8 | — | — | — | — | — | — | — | — | |
| Q2Bembedding dimension (d)=16002021.10 | 21.1 | — | — | — | 31.1 | 45.4 | 11.9 | — | — | — | 41.3 | 9.9 | 7.2 | 21.9 | 13.3 | 8.1 | — | — | — | — | — | — | — | — | |
| BETAEembedding dimension (d)=4002021.10 | 20.9 | — | — | — | 28.8 | 42.5 | 12.4 | — | — | — | 39 | 10.9 | 10 | 22.4 | 12.6 | 9.7 | — | — | — | — | — | — | — | — | |
| GQEembedding dimension (d)=16002021.10 | 16.6 | — | — | — | 23.6 | 35.7 | 8.4 | — | — | — | 35.2 | 7.4 | 5.5 | 16.7 | 10.9 | 5.8 | — | — | — | — | — | — | — | — | |
| ConE+CaQR(R)base_model=ConE, embedding_type=relation-induced embedding2024.06 | 6.85 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 6.84 | 10.44 | 7.66 | 4.49 | 4.79 | |
| ConEbase_model=ConE2024.06 | 6.31 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.78 | 9.57 | 7.95 | 3.85 | 4.41 | |
| ConE+CaQR(S)base_model=ConE, embedding_type=structure embedding2024.06 | 6.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 6.25 | 9.62 | 7.6 | 3.97 | 4.03 | |
| ConE+CaQRbase_model=ConE, embedding_type=structure and relation-induced embedding2024.06 | 5.93 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.24 | 8.91 | 8.07 | 3.51 | 3.9 | |
| BetaE+CaQRbase_model=BetaE, embedding_type=structure and relation-induced embedding2024.06 | 5.88 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.49 | 8.24 | 8.01 | 3.77 | 3.91 | |
| BetaE+CaQR(R)base_model=BetaE, embedding_type=relation-induced embedding2024.06 | 5.81 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.42 | 8.3 | 7.83 | 3.6 | 3.89 | |
| BetaE+CaQR(S)base_model=BetaE, embedding_type=structure embedding2024.06 | 5.76 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.29 | 8.35 | 7.77 | 3.51 | 3.9 | |
| BetaEbase_model=BetaE2024.06 | 5.54 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 5.19 | 7.94 | 7.44 | 3.6 | 3.55 | |
| nPUGraphptb_rate=0.1, Category=Positive-Unlabeled learning method2023.06 | 0.343 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.328 | 0.535 | 0.221 | — | — | — | — | — | |
| RotatEptb_rate=0.1, Category=Knowledge graph embedding methods2023.06 | 0.324 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.292 | 0.48 | 0.199 | — | — | — | — | — | |
| CompGCNptb_rate=0.1, Category=Graph neural network methods on KG2023.06 | 0.324 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.293 | 0.475 | 0.203 | — | — | — | — | — | |
| PERM2021.10 | 0.306 | 0.52 | 0.286 | 0.216 | 0.361 | 0.49 | 0.305 | 0.128 | 0.212 | 0.239 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ComplExptb_rate=0.1, Category=Knowledge graph embedding methods2023.06 | 0.305 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.276 | 0.459 | 0.186 | — | — | — | — | — | |
| SANSptb_rate=0.1, Category=Negative sampling method2023.06 | 0.301 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.271 | 0.453 | 0.182 | — | — | — | — | — | |
| PUDAptb_rate=0.1, Category=Positive-Unlabeled learning method2023.06 | 0.298 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.271 | 0.443 | 0.185 | — | — | — | — | — | |
| DistMultptb_rate=0.1, Category=Knowledge graph embedding methods2023.06 | 0.297 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.271 | 0.446 | 0.185 | — | — | — | — | — | |
| CQD2021.10 | 0.29 | 0.512 | 0.288 | 0.221 | 0.352 | 0.457 | 0.284 | 0.129 | 0.249 | 0.121 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TransEptb_rate=0.1, Category=Knowledge graph embedding methods2023.06 | 0.284 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.245 | 0.441 | 0.144 | — | — | — | — | — | |
| TransRptb_rate=0.1, Category=Knowledge graph embedding methods2023.06 | 0.28 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.246 | 0.434 | 0.153 | — | — | — | — | — | |
| Q2B2021.10 | 0.251 | 0.467 | 0.24 | 0.186 | 0.324 | 0.453 | 0.239 | 0.05 | 0.108 | 0.193 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GQE2021.10 | 0.221 | 0.404 | 0.214 | 0.147 | 0.262 | 0.39 | 0.164 | 0.087 | 0.162 | 0.155 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BQE2021.10 | 0.218 | 0.455 | 0.122 | 0.102 | 0.232 | 0.459 | 0.141 | 0.224 | 0.124 | 0.101 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UKGEptb_rate=0.1, Category=Uncertain knowledge graph embedding method2023.06 | 0.204 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.231 | 0.411 | 0.11 | — | — | — | — | — | |
| RGCNptb_rate=0.1, Category=Graph neural network methods on KG2023.06 | 0.177 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.169 | 0.317 | 0.097 | — | — | — | — | — | |
| NSCachingptb_rate=0.1, Category=Negative sampling method2023.06 | 0.167 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.153 | 0.305 | 0.08 | — | — | — | — | — | |
| CSProm-KGBackbone=BERT2025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.358 | 0.538 | — | — | — | — | — | — | |
| GPT-42025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.42 | 0.565 | — | — | — | — | — | — | |
| KICGPTBackbone=GPT-3.52025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.412 | 0.554 | — | — | — | — | — | — | |
| KRLMStage=Pre-training2025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.381 | 0.554 | — | — | — | — | — | — | |
| KRLMStage=End-to-end training2025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.394 | 0.568 | — | — | — | — | — | — | |
| MKGL2025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.41 | 0.591 | — | — | — | — | — | — | |
| MOTIFStage=Pre-training2025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.357 | 0.55 | — | — | — | — | — | — | |
| NBFNet2022.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.415 | — | — | — | — | — | — | — | |
| RulEKGE Model=RotatE2022.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.362 | — | — | — | — | — | — | — | |
| TRIXStage=Pre-training2025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.366 | 0.559 | — | — | — | — | — | — | |
| ULTRAStage=Pre-training2025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 0.368 | 0.564 | — | — | — | — | — | — |