Knowledge Graph Reasoning on WN18RR
67MRRNCRL
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| NCRL2022.10 | 67 | 56.3 | — | 85 | |
| RulEvariant=emb & rule., backbone=RotatE2022.10 | 51.9 | 47.5 | 53.8 | 60.5 | |
| RNNLogic+variant=with emb., source=Tuned rerun2022.10 | 51.6 | 46.9 | 52.8 | 60.4 | |
| RulEvariant=rule.2022.10 | 51.4 | 47.3 | 53.3 | 59.7 | |
| RNNLogic+variant=with emb., source=Original paper2022.10 | 51.3 | 47.1 | 53.2 | 59.7 | |
| Naive Combinationsource=Rerun2022.10 | 51.2 | 46.9 | 53.7 | 59.7 | |
| RulEvariant=emb., backbone=RotatE2022.10 | 51 | 41.4 | 57.3 | 68.2 | |
| RNNLogic+variant=w/o emb., source=Tuned rerun2022.10 | 50.2 | 46.1 | 52.2 | 58.5 | |
| RNNLogic+variant=w/o emb., source=Original paper2022.10 | 48.9 | 45.3 | 50.6 | 56.3 | |
| DRUMsource=Original paper2022.10 | 48.6 | 42.5 | 51.3 | 58.6 | |
| RotatEsource=Rerun2022.10 | 47.6 | 43.1 | 49.2 | 56.2 | |
| TuckERsource=Original paper2022.10 | 47 | 44.3 | 48.2 | 52.6 | |
| ComplExsource=Original paper2022.10 | 44 | 41 | 46 | 51 | |
| Neural-LPsource=Original paper2022.10 | 43.5 | 37.1 | 43.4 | 56.6 | |
| DistMultsource=Original paper2022.10 | 43 | 39 | 44 | 49 | |
| ConvEsource=Original paper2022.10 | 43 | 40 | 44 | 52 | |
| RulEvariant=emb., backbone=TransE2022.10 | 24.2 | 6.7 | 37.8 | 52.6 | |
| TransEsource=Rerun2022.10 | 22.2 | 1.2 | 39.9 | 53 | |
| PathRanksource=Original paper2022.10 | 18.9 | 17.1 | 20 | 22.5 |