Entity Alignment on DBP15K FR-EN
0.996Hits@1MEAformer
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MEAformerIterative=true, Surface Forms=w/ SF, Para.=14.2M2022.12 | 0.996 | 1 | 0.998 | |
| AgentEAMethod category=Large Language Models-based2026.04 | 0.996 | 1 | 0.998 | |
| MCLEA-NEvaluation Protocol=Unsupervised, Pseudo-alignment seed generation=Name-based (-N)2022.09 | 0.995 | 1 | 0.997 | |
| ProLEAMethod category=Large Language Models-based2026.04 | 0.993 | 1 | 0.997 | |
| MEAformerPara.=14.2M, Surface Forms (SF)=with, Non-iterative=true2022.12 | 0.991 | 1 | 0.995 | |
| MEAformerIterative=false, Surface Forms=w/ SF, Para.=14.2M2022.12 | 0.991 | 1 | 0.995 | |
| MCLEA-VEvaluation Protocol=Unsupervised, Pseudo-alignment seed generation=Visual-based (-V)2022.09 | 0.99 | 1 | 0.994 | |
| ChatEABackbone=GPT-42026.04 | 0.99 | — | 0.995 | |
| EA-Agent2026.04 | 0.99 | 0.997 | 0.996 | |
| ChatEAMethod category=Large Language Models-based2026.04 | 0.99 | 1 | 0.995 | |
| SEUEvaluation Protocol=Unsupervised2022.09 | 0.988 | 0.999 | 0.992 | |
| BERT-INTPara.=>110M, Surface Forms (SF)=with, Non-iterative=true2022.12 | 0.987 | 0.992 | 0.99 | |
| BERT-INT2026.04 | 0.987 | 0.992 | 0.99 | |
| Seg-AlignBackbone=GPT-3.52026.04 | 0.987 | — | 0.987 | |
| Seg-AlignMethod category=Large Language Models-based2026.04 | 0.987 | — | — | |
| SCMEAMethod category=Knowledge Representation Learning-based2026.04 | 0.982 | 0.998 | 0.988 | |
| EASYEvaluation Protocol=Unsupervised2022.09 | 0.98 | 0.998 | 0.99 | |
| SDEAPara.=>110M, Surface Forms (SF)=with, Non-iterative=true2022.12 | 0.969 | 0.995 | 0.98 | |
| SDEA2026.04 | 0.966 | 0.995 | 0.98 | |
| AdaCoAgentEAMethod category=Large Language Models-based2026.04 | 0.964 | 0.982 | 0.971 | |
| RREAEvaluation Protocol=Unsupervised2022.09 | 0.963 | 0.992 | — | |
| TEA2026.04 | 0.958 | 0.997 | 0.97 | |
| LLMEABackbone=ERNIE2026.04 | 0.957 | — | 0.957 | |
| LLMEAMethod category=Large Language Models-based2026.04 | 0.957 | 0.977 | — | |
| Dual-AMNMethod category=Knowledge Representation Learning-based2026.04 | 0.954 | 0.994 | 0.97 | |
| MRAEAEvaluation Protocol=Unsupervised2022.09 | 0.95 | 0.97 | — | |
| RAEAEnsemble strategy=pre-weighted strategy2025.12 | 0.9497 | 0.996 | 0.968 | |
| Ours (sparse)L=102020.01 | 0.9334 | 0.9603 | — | |
| EasyEA*Method category=Large Language Models-based2026.04 | 0.927 | 0.998 | — | |
| Ours (sparse)L=102020.01 | 0.9195 | 0.9528 | — | |
| AttrGNN2025.12 | 0.9185 | 0.9777 | 0.91 | |
| NMN2025.12 | 0.902 | 0.967 | 0.864 | |
| GMNN2020.01 | 0.8938 | 0.9525 | — | |
| HGCN-JEjoint embeddings=true, initialization=machine-translated entity names, layer-wise highway gates=true2019.09 | 0.8916 | 0.9611 | — | |
| HGCN-PEinitialization=machine-translated entity names, layer-wise highway gates=true2019.09 | 0.8873 | 0.9552 | — | |
| RDGCN2020.01 | 0.8864 | 0.9572 | — | |
| RDGCN2025.12 | 0.8864 | 0.9572 | 0.908 | |
| RDGCN2026.04 | 0.886 | 0.957 | 0.911 | |
| GMNN2020.01 | 0.8818 | 0.9475 | — | |
| PMFIteration strategy=Iterative2024.07 | 0.879 | — | 0.921 | |
| UMAEA*Iteration strategy=Iterative2024.07 | 0.873 | — | 0.917 | |
| RDGCNMethod category=Knowledge Representation Learning-based2026.04 | 0.873 | 0.95 | 0.901 | |
| RAEAEnsemble strategy=average pooling2025.12 | 0.8582 | 0.9749 | 0.902 | |
| PMFIteration strategy=Non-iterative2024.07 | 0.85 | — | 0.898 | |
| MEAformer*Iteration strategy=Iterative2024.07 | 0.845 | — | 0.894 | |
| Ours (sparse)L=02020.01 | 0.8336 | 0.9603 | — | |
| UMAEA*Iteration strategy=Non-iterative2024.07 | 0.832 | — | 0.887 | |
| RREAsupervision=semi-supervised2020.08 | 0.827 | 0.966 | 0.881 | |
| Ours (sparse)L=02020.01 | 0.8216 | 0.9528 | — | |
| MEAformer*Iteration strategy=Non-iterative2024.07 | 0.818 | — | 0.877 | |
| MCLEAIteration strategy=Iterative2024.07 | 0.808 | — | 0.867 | |
| EVAIterative Learning (IL)=true2020.09 | 0.793 | 0.942 | 0.847 | |
| MRAEAsupervision=semi-supervised2020.08 | 0.781 | 0.935 | 0.849 | |
| MCLEAIteration strategy=Non-iterative2024.07 | 0.772 | — | 0.841 | |
| MEAformerIterative=false, Surface Forms=w/o SF, Para.=13.7M2022.12 | 0.77 | 0.961 | 0.841 | |
| JEANSVariant=EDL2020.05 | 0.769 | 0.94 | 0.827 | |
| JEANSVariant=SFM2020.05 | 0.766 | 0.939 | 0.814 | |
| EVAIteration strategy=Iterative2024.07 | 0.765 | — | 0.831 | |
| RAEAEnsemble strategy=SVM2025.12 | 0.7623 | 0.9365 | 0.827 | |
| Ψθ₁ = MLPL=02020.01 | 0.7607 | 0.9154 | — | |
| Dual-AMN2025.12 | 0.756 | 0.948 | 0.827 | |
| Ψθ₁ = MLPL=02020.01 | 0.7489 | 0.9057 | — | |
| HyperKA (semi)dimension=75, learning_type=semi-supervised2020.10 | 0.741 | — | 0.813 | |
| RREAsupervision=basic2020.08 | 0.739 | 0.946 | 0.816 | |
| EVAIteration strategy=Non-iterative2024.07 | 0.719 | — | 0.792 | |
| EVAIterative Learning (IL)=false2020.09 | 0.715 | 0.936 | 0.795 | |
| MultiKE2025.12 | 0.7143 | 0.7608 | 0.733 | |
| TransEdge2025.12 | 0.7101 | 0.9414 | 0.796 | |
| TransEdgesupervision=semi-supervised, translation-based=true2020.08 | 0.71 | 0.941 | 0.796 | |
| TransEdgedimension=75, learning_type=semi-supervised2020.10 | 0.71 | — | 0.796 | |
| MSNEAIteration strategy=Non-iterative2024.07 | 0.686 | — | 0.771 | |
| NAEA2020.01 | 0.6732 | 0.8943 | — | |
| NAEAsupervision=semi-supervised2020.08 | 0.673 | 0.894 | 0.752 | |
| NAEAdimension=75, learning_type=semi-supervised2020.10 | 0.673 | — | 0.752 | |
| NAEAIterative Learning (IL)=true2020.09 | 0.673 | 0.894 | 0.752 | |
| NAEA2020.05 | 0.673 | 0.894 | 0.752 | |
| MRAEAsupervision=basic2020.08 | 0.666 | 0.912 | 0.765 | |
| BootEA2019.09 | 0.653 | 0.8744 | — | |
| BOOTEA2020.01 | 0.653 | 0.8744 | — | |
| BootEAsupervision=semi-supervised, translation-based=true2020.08 | 0.653 | 0.874 | 0.731 | |
| BootEAdimension=75, learning_type=semi-supervised2020.10 | 0.653 | — | 0.731 | |
| BOOTEAIterative Learning (IL)=true2020.09 | 0.653 | 0.874 | 0.731 | |
| BootEA2020.05 | 0.653 | 0.874 | 0.731 | |
| BootEA2025.12 | 0.653 | 0.8744 | 0.731 | |
| BootEAMethod category=Knowledge Representation Learning-based2026.04 | 0.653 | 0.874 | 0.731 | |
| TransEdgesupervision=basic, translation-based=true2020.08 | 0.649 | 0.921 | 0.746 | |
| MMEAIterative Learning (IL)=true2020.09 | 0.635 | 0.878 | — | |
| RKDEA2025.12 | 0.622 | 0.912 | 0.721 | |
| HyperKADimensions=75, 75, 752020.10 | 0.597 | 0.891 | 0.704 | |
| GMNIterative Learning (IL)=false2020.09 | 0.596 | 0.876 | 0.679 | |
| GMN2020.05 | 0.596 | 0.876 | 0.679 | |
| GCN-JEjoint embeddings=true, initialization=machine-translated entity names2019.09 | 0.583 | 0.8634 | — | |
| MSNEAIteration strategy=Iterative2024.07 | 0.583 | — | 0.672 | |
| MMEAIteration strategy=Non-iterative2024.07 | 0.557 | — | 0.643 | |
| AliNetfull version=true2019.11 | 0.552 | 0.852 | 0.657 | |
| AliNetsupervision=basic2020.08 | 0.552 | 0.852 | 0.657 | |
| AliNetDimensions=500, 400, 3002020.10 | 0.552 | 0.852 | 0.657 | |
| ALINETIterative Learning (IL)=false2020.09 | 0.552 | 0.852 | 0.657 | |
| AliNet2020.05 | 0.552 | 0.852 | 0.657 | |
| AliNet2025.12 | 0.5495 | 0.8313 | 0.645 |