Graph Classification on MUTAG
98.8AccuracyCDAT
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| CDATevaluation_protocol=100 runs of 10-fold cross-validation2026.06 | 98.8 | — | — | — | — | — | — | — | |
| PR-MPNN2024.05 | 98.4 | — | — | — | — | — | — | — | |
| IPR-MPNN2024.05 | 98 | — | — | — | — | — | — | — | |
| ISP-GNN_DegreeEvaluation protocol=10-fold cross-validation2026.03 | 97.9 | — | — | — | — | — | — | — | |
| ID-GNNEvaluation protocol=10-fold cross-validation2026.03 | 97.8 | — | — | — | — | — | — | — | |
| ISP-GNN†Evaluation protocol=10-fold cross-validation2026.03 | 97.5 | — | — | — | — | — | — | — | |
| AC-GINEvaluation protocol=10-fold cross-validation2026.03 | 96.8 | — | — | — | — | — | — | — | |
| ISP-GNN_OnionEvaluation protocol=10-fold cross-validation2026.03 | 96.8 | — | — | — | — | — | — | — | |
| ISP-GNN_CoreEvaluation protocol=10-fold cross-validation2026.03 | 96.7 | — | — | — | — | — | — | — | |
| SOTA2024.10 | 96.66 | — | — | — | — | — | — | — | |
| ETevaluation_protocol=100 runs of 10-fold cross-validation2026.06 | 96.6 | — | — | — | — | — | — | — | |
| CCIN2026.05 | 96.4 | — | — | — | — | — | — | — | |
| BEC-GINEvaluation protocol=10-fold cross-validation2026.03 | 96.1 | — | — | — | — | — | — | — | |
| KP-GINEvaluation protocol=10-fold cross-validation2026.03 | 95.6 | — | — | — | — | — | — | — | |
| GIN-AK+Evaluation protocol=10-fold cross-validation2026.03 | 95 | — | — | — | — | — | — | — | |
| NGRAMCNN2018.11 | 94.99 | — | — | — | — | — | — | — | |
| PXGL-GNN2025.12 | 94.87 | — | — | — | — | — | — | — | |
| AutoGELType=AutoGNN2021.12 | 94.74 | — | — | — | — | — | — | — | |
| GraphSNNEvaluation protocol=10-fold cross-validation2026.03 | 94.7 | — | — | — | — | — | — | — | |
| HMH2026.05 | 94.5 | — | — | — | — | — | — | — | |
| GIC2018.11 | 94.44 | — | — | — | — | — | — | — | |
| CIN++Type=Topological neural network2023.06 | 94.4 | — | — | — | — | — | — | — | |
| CIN++2024.05 | 94.4 | — | — | — | — | — | — | — | |
| (s)GIN-0neighborhood enlargement=true, epsilon=02019.05 | 94.14 | — | — | — | — | — | — | — | |
| CANType=Topological neural network2023.06 | 94.1 | — | — | — | — | — | — | — | |
| CAN2024.05 | 94.1 | — | — | — | — | — | — | — | |
| GraphFMfine-tuning=MLP fine-tuning, decoder=learnable cross-attention decoder2024.07 | 93.8 | — | — | — | — | — | — | — | |
| UnionGINEvaluation protocol=10-fold cross-validation2026.03 | 93.6 | — | — | — | — | — | — | — | |
| GCBM-EProtocol=10-fold cross-validation, Representation=Dense embeddings (Transformer)2026.03 | 93.6 | — | — | — | — | — | — | — | |
| VEPMLearning Paradigm=Supervised/Semi-supervised2024.08 | 93.6 | — | — | — | — | — | — | — | |
| StructPoolevaluation_protocol=10-fold cross-validation2021.07 | 93.59 | — | — | — | — | — | — | — | |
| GMCL2026.04 | 93.58 | — | — | — | — | — | — | — | |
| GCBMProtocol=10-fold cross-validation, Representation=One-hot2026.03 | 93.5 | — | — | — | — | — | — | — | |
| (s)GIN-epsilonneighborhood enlargement=true2019.05 | 93.47 | — | — | — | — | — | — | — | |
| SAGNN2025.12 | 93.24 | — | — | — | — | — | — | — | |
| GSNEvaluation protocol=10-fold cross-validation2026.03 | 93.1 | — | — | — | — | — | — | — | |
| CH-CL2026.04 | 93.02 | — | — | — | — | — | — | — | |
| ChebNetDiffusion Mechanism=KHG, K=122026.06 | 92.98 | — | — | — | — | — | — | — | |
| ALL-IN2026.05 | 92.9 | — | — | — | — | — | — | — | |
| ALL-INpropagated covariance operators=full2026.05 | 92.9 | — | — | — | — | — | — | — | |
| EGIN-Cepsilon parameter=not retained, validation=10-fold cross-validation2025.12 | 92.8 | — | — | — | — | — | — | — | |
| GINEvaluation protocol=10-fold cross-validation2026.03 | 92.8 | — | — | — | — | — | — | — | |
| NC-GNNEvaluation protocol=10-fold cross-validation2026.03 | 92.8 | — | — | — | — | — | — | — | |
| CINType=Topological neural network2023.06 | 92.7 | — | — | — | — | — | — | — | |
| CIN2024.05 | 92.7 | — | — | — | — | — | — | — | |
| TopNets2026.05 | 92.7 | — | — | — | — | — | — | — | |
| CINPeak GPU Memory (MB)=312026.05 | 92.7 | — | — | — | — | — | — | — | |
| SAN2025.12 | 92.65 | — | — | — | — | — | — | — | |
| PSCN2018.05 | 92.63 | — | — | — | — | — | — | — | |
| PSCNk=10, node_attribute=normalized degree, time (seconds)=32016.05 | 92.63 | — | — | — | — | — | — | — | |
| PSCN2018.11 | 92.63 | — | — | — | — | — | — | — | |
| PSCNInference Setting=Inductive, Evaluation Protocol=10-fold cross-validation2019.09 | 92.63 | — | — | — | — | — | — | — | |
| PSCNedge types=annotated only2018.11 | 92.6 | — | — | — | — | — | — | — | |
| PATCHYSANMethod category=GNN2018.10 | 92.6 | — | — | — | — | — | — | — | |
| PATCHYSANType=Manual GNNs2021.12 | 92.6 | — | — | — | — | — | — | — | |
| SATProtocol=10-fold cross-validation, Group=Group 52026.03 | 92.6 | — | — | — | — | — | — | — | |
| DisenSemiLearning Paradigm=Supervised/Semi-supervised2024.08 | 92.6 | — | — | — | — | — | — | — | |
| CTQWformer2026.05 | 92.54 | — | — | — | — | — | — | — | |
| CTQWformer2026.05 | 92.54 | — | — | — | — | — | — | — | |
| GIN + TanhBackbone=GIN, Activation=Tanh2024.07 | 92.5 | — | — | — | — | — | — | — | |
| GIN + ELUBackbone=GIN, Activation=ELU2024.07 | 92.5 | — | — | — | — | — | — | — | |
| EGINepsilon parameter=not retained, validation=10-fold cross-validation2025.12 | 92.5 | — | — | — | — | — | — | — | |
| DropGINEvaluation protocol=10-fold cross-validation2026.03 | 92.5 | — | — | — | — | — | — | — | |
| ALL-INprops=02026.05 | 92.5 | — | — | — | — | — | — | — | |
| ALL-INpropagated covariance operators=02026.05 | 92.5 | — | — | — | — | — | — | — | |
| G2N2Peak GPU Memory (MB)=1892026.05 | 92.5 | — | — | — | — | — | — | — | |
| GSNType=Graph neural network2023.06 | 92.2 | — | — | — | — | — | — | — | |
| GSN2024.05 | 92.2 | — | — | — | — | — | — | — | |
| GSN2026.05 | 92.2 | — | — | — | — | — | — | — | |
| GIN + GRANOLAevaluation_protocol=100 runs of 10-fold cross-validation2026.06 | 92.2 | — | — | — | — | — | — | — | |
| FGSD2023.05 | 92.12 | — | — | — | — | — | — | — | |
| FGSDnode features=false2019.02 | 92.1 | — | — | — | — | — | — | — | |
| FGSD2018.11 | 92.1 | — | — | — | — | — | — | — | |
| GIN + DIGRAFBackbone=GIN, Activation=DIGRAF, Adaptive=true2024.07 | 92.1 | — | — | — | — | — | — | — | |
| GRDLProtocol=10-fold cross-validation, Group=Group 42026.03 | 92.1 | — | — | — | — | — | — | — | |
| DGCLLearning Paradigm=Unsupervised2024.08 | 92.1 | — | — | — | — | — | — | — | |
| GRDLevaluation_protocol=100 runs of 10-fold cross-validation2026.06 | 92.1 | — | — | — | — | — | — | — | |
| GIN + MedianBackbone=GIN, Activation=Median2024.07 | 92 | — | — | — | — | — | — | — | |
| GIN + GRELUBackbone=GIN, Activation=GRELU2024.07 | 92 | — | — | — | — | — | — | — | |
| GIN + DIGRAF (W/O ADAP.)Backbone=GIN, Activation=DIGRAF, Adaptive=false2024.07 | 92 | — | — | — | — | — | — | — | |
| FractalGCLEvaluation Protocol=10-fold CV2025.05 | 91.71 | — | — | — | — | — | — | — | |
| GIN + PReLUBackbone=GIN, Activation=PReLU2024.07 | 91.7 | — | — | — | — | — | — | — | |
| sCWNrPeak GPU Memory (MB)=222026.05 | 91.7 | — | — | — | — | — | — | — | |
| OT-GNNProtocol=10-fold cross-validation, Group=Group 32026.03 | 91.6 | — | — | — | — | — | — | — | |
| GIN + GraphNormevaluation_protocol=100 runs of 10-fold cross-validation2026.06 | 91.6 | — | — | — | — | — | — | — | |
| PSCNk=5, time (seconds)=22016.05 | 91.58 | — | — | — | — | — | — | — | |
| DDMProtocol=Unsupervised representation learning2023.06 | 91.51 | — | — | — | — | — | — | — | |
| ARMA2019.01 | 91.5 | — | — | — | — | — | — | — | |
| GIN + MaxoutBackbone=GIN, Activation=Maxout2024.07 | 91.5 | — | — | — | — | — | — | — | |
| LinearKernel Type=Classical Linear2025.11 | 91.49 | — | 89.83 | — | — | — | — | — | |
| RBFKernel Type=Classical RBF2025.11 | 91.49 | — | 89.83 | — | — | — | — | — | |
| QRBFKernel Type=Quantum RBF2025.11 | 91.49 | — | 89.83 | — | — | — | — | — | |
| QAmpKernel Type=Quantum Amplitude2025.11 | 91.49 | — | 89.83 | — | — | — | — | — | |
| GIN + IdentityBackbone=GIN, Activation=Identity2024.07 | 91.4 | — | — | — | — | — | — | — | |
| (s)GCNevaluation=10-cross validation, continuous node attribute=false, batch normalization=false2019.05 | 91.39 | — | — | — | — | — | — | — | |
| ICL2025.12 | 91.34 | — | — | — | — | — | — | — | |
| GCMAE2025.12 | 91.3 | — | — | — | — | — | — | — | |
| Graph-JEPAPre-training Type=Self-predictive2023.09 | 91.25 | — | — | — | — | — | — | — | |
| InfoGCLLearning paradigm=Unsupervised Approaches2021.10 | 91.2 | — | — | — | — | — | — | — | |
| InfoGCL2025.12 | 91.2 | — | — | — | — | — | — | — |