Link Property Prediction On Ogbl Wikikg2
評価指標
Ext. data
Number of params
Test MRR
Validation MRR
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
比較表
| モデル名 | Ext. data | Number of params | Test MRR | Validation MRR |
|---|---|---|---|---|
| interht-knowledge-graph-embeddings-by | No | 148000738 | 0.7257 ± 0.0018 | 0.7370 ± 0.0022 |
| relation-prediction-as-an-auxiliary-training | - | 250167400 | 0.6364 | 0.6594 |
| complex-embeddings-for-simple-link-prediction | No | 250113900 | 0.3804 ± 0.0022 | 0.3534 ± 0.0052 |
| rotate-knowledge-graph-embedding-by | No | 250087150 | 0.2530 ± 0.0034 | 0.2250 ± 0.0035 |
| complex-embeddings-for-simple-link-prediction | No | 1250569500 | 0.4027 ± 0.0027 | 0.3759 ± 0.0016 |
| モデル 6 | No | 500763337 | 0.5794 ± 0.0020 | 0.6045 ± 0.0024 |
| embedding-entities-and-relations-for-learning | No | 1250569500 | 0.3729 ± 0.0045 | 0.3506 ± 0.0042 |
| rot-pro-modeling-transitivity-by-projection | No | 1000669602 | 0.5602 ± 0.0016 | 0.5740 ± 0.0008 |
| stargraph-a-coarse-to-fine-representation | No | 86762146 | 0.7201 ± 0.0011 | 0.7288 ± 0.0008 |
| interht-knowledge-graph-embeddings-by | No | 156332770 | 0.7293 ± 0.0018 | 0.7391 ± 0.0023 |
| embedding-entities-and-relations-for-learning | No | 250113900 | 0.3447 ± 0.0082 | 0.3150 ± 0.0088 |
| autoweird-weird-translational-scoring | No | 19215402 | 0.7353 ± 0.0006 | 0.7362 ± 0.0006 |
| interht-knowledge-graph-embeddings-by | No | 19215402 | 0.6779 ± 0.0018 | 0.6893 ± 0.0015 |
| knowledge-graph-embedding-with-3d-compound | No | 750662700 | 0.7006 ± 0.0011 | 0.7175 ± 0.0015 |
| モデル 15 | No | 1250569500 | 0.4256 ± 0.0030 | 0.4272 ± 0.0030 |
| autokge-searching-scoring-functions-for | No | 500227800 | 0.5458 ± 0.0052 | 0.5510 ± 0.0063 |
| pairre-knowledge-graph-embeddings-via-paired | No | 500334800 | 0.5208 ± 0.0027 | 0.5423 ± 0.0020 |
| transerr-translation-based-knowledge-graph | No | 500441802 | 0.6359 ± 0.0020 | 0.6518 ± 0.0012 |
| モデル 19 | No | 250167400 | 0.6392 ± 0.0045 | 0.6561 ± 0.0070 |
| nodepiece-compositional-and-parameter | No | 6860602 | 0.5703 ± 0.0035 | 0.5806 ± 0.0047 |
| モデル 21 | Yes | 1927395330 | 0.7305 ± 0.0010 | 0.7442 ± 0.0006 |
| relation-prediction-as-an-auxiliary-training | - | 500334800 | 0.6481 | 0.6701 |
| モデル 23 | No | 6831201 | 0.6711 ± 0.0045 | 0.6796 ± 0.0070 |
| trans-transition-based-knowledge-graph | No | 19215402 | 0.6882 ± 0.0019 | 0.6988 ± 0.0006 |
| rot-pro-modeling-transitivity-by-projection | No | 1000669602 | 0.4277 ± 0.0008 | 0.4174 ± 0.0058 |
| モデル 26 | No | 250113900 | 0.2622 ± 0.0045 | 0.2465 ± 0.0020 |
| a-net-a-scalable-path-based-reasoning-1 | No | 6831201 | 0.6711 ± 0.0045 | 0.6796 ± 0.0070 |
| モデル 28 | No | 7289002 | 0.6582 ± 0.0020 | 0.6616 ± 0.0018 |
| rotate-knowledge-graph-embedding-by | No | 1250435750 | 0.4332 ± 0.0025 | 0.4353 ± 0.0028 |
| relation-aware-ensemble-learning-for | No | 2176767622 | 0.7392 ± 0.0011 | 0.7509 ± 0.0009 |