Node Classification on Cora
99.99AccuracyR-DCR
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| R-DCRRelational=true, Interpretability=Yes, Rules=Learnt2023.08 | 99.99 | — | — | — | — | |
| CGRLBackbone=GCN2026.03 | 99.14 | — | — | — | — | |
| CGRLBackbone=GAT2026.03 | 98.56 | — | — | — | — | |
| CIA-LRABackbone=GAT2026.03 | 97.89 | — | — | — | — | |
| CIA-LRABackbone=GCN2026.03 | 97.43 | — | — | — | — | |
| CaNetBackbone=GAT2026.03 | 97.3 | — | — | — | — | |
| CaNetBackbone=GCN2026.03 | 96.12 | — | — | — | — | |
| CNAArchitecture=GraphSAGE2024.12 | 94.18 | — | — | — | — | |
| MixupBackbone=GAT2026.03 | 92.94 | — | — | — | — | |
| MixupBackbone=GCN2026.03 | 92.77 | — | — | — | — | |
| PromptGFM*Backbone=GPT-4o2026.03 | 92.42 | — | — | — | — | |
| DANNBackbone=GAT2026.03 | 92.4 | — | — | — | — | |
| CoralBackbone=GAT2026.03 | 91.82 | — | — | — | — | |
| EERMBackbone=GAT2026.03 | 91.8 | — | — | — | — | |
| SRGNNBackbone=GAT2026.03 | 91.77 | — | — | — | — | |
| IRMBackbone=GAT2026.03 | 91.63 | — | — | — | — | |
| ERMBackbone=GAT2026.03 | 91.1 | — | — | — | — | |
| GroupDROBackbone=GAT2026.03 | 90.54 | — | — | — | — | |
| DAG+NodeFormerBackbone=NodeFormer, Extension=DAG Attention2022.10 | 90.49 | — | — | — | — | |
| SSPArchitecture=SSP2024.12 | 90.16 | — | — | — | — | |
| AutoGELType=AutoGNN2021.12 | 89.89 | — | — | — | — | |
| HiGNN2024.03 | 89.72 | — | — | — | — | |
| SANEType=AutoGNN2021.12 | 89.26 | — | — | — | — | |
| FDiff-scaleBase Model=Plain Linear2020.10 | 89.05 | — | — | — | — | |
| SIGMA2023.05 | 88.96 | — | — | — | — | |
| DANCERatio=8%2026.01 | 88.87 | — | — | — | — | |
| NodeFormerBackbone=NodeFormer2022.10 | 88.8 | — | — | — | — | |
| DiagChebyT4SDSheaf Type=Diagonal, Polynomial=Chebyshev, Order=T42025.11 | 88.79 | — | — | — | — | |
| UGTpre-training=true2023.08 | 88.74 | — | — | — | — | |
| AutoscaleBase Model=Linear2020.10 | 88.73 | — | — | — | — | |
| LHS2023.12 | 88.71 | — | — | — | — | |
| DiagChebyT1SDSheaf Type=Diagonal, Polynomial=Chebyshev, Order=T12025.11 | 88.67 | — | — | — | — | |
| FDiff-scaleBase Model=Linear2020.10 | 88.66 | — | — | — | — | |
| AutoscaleBase Model=Plain Linear2020.10 | 88.62 | — | — | — | — | |
| PDE-GCNMLayers=162021.08 | 88.6 | — | — | — | — | |
| GREAD-BS2022.11 | 88.57 | — | — | — | — | |
| PDE-GCNDLayers=162021.08 | 88.51 | — | — | — | — | |
| You et al. [2020]Type=AutoGNN2021.12 | 88.5 | — | — | — | — | |
| NLGAT2020.05 | 88.5 | — | — | — | — | |
| NLGNN2023.12 | 88.5 | — | — | — | — | |
| M-DESIGN2025.07 | 88.5 | — | — | — | — | |
| SOTA2020.10 | 88.49 | — | — | — | — | |
| GCNIILayers=642021.08 | 88.49 | — | — | — | — | |
| GCNIIsource=Chen et al. (2020), Number of layers=642021.10 | 88.49 | — | — | — | — | |
| GREAD-ST2022.11 | 88.47 | — | — | — | — | |
| GraphNASType=AutoGNN2021.12 | 88.4 | — | — | — | — | |
| GAT2020.05 | 88.4 | — | — | — | — | |
| GREAD-F2022.11 | 88.39 | — | — | — | — | |
| GCNII2022.06 | 88.37 | — | — | — | — | |
| GCNIItrials=102022.05 | 88.37 | — | — | — | — | |
| GCNIILayers=Best across layers2021.02 | 88.37 | — | — | — | — | |
| Ordered GNN2023.02 | 88.37 | — | — | — | — | |
| GCNIIVariant=best architecture2023.02 | 88.37 | — | — | — | — | |
| GCNII2022.11 | 88.37 | — | — | — | — | |
| GCNII2023.08 | 88.37 | — | — | — | — | |
| GCNII2024.05 | 88.37 | — | — | — | — | |
| GCNII2023.05 | 88.37 | — | — | — | — | |
| GCNII2025.11 | 88.37 | — | — | — | — | |
| GCNII2023.05 | 88.37 | — | — | — | — | |
| GloGNN++trials=102022.05 | 88.33 | — | — | — | — | |
| GloGNN2024.05 | 88.33 | — | — | — | — | |
| GloGNN++2023.05 | 88.33 | — | — | — | — | |
| GloGNN2023.05 | 88.33 | — | — | — | — | |
| GloGNNtrials=102022.05 | 88.31 | — | — | — | — | |
| GloGNN2022.11 | 88.31 | — | — | — | — | |
| GREAD-Z2022.11 | 88.31 | — | — | — | — | |
| GloGNN2023.08 | 88.31 | — | — | — | — | |
| GloGNN2023.05 | 88.31 | — | — | — | — | |
| GloGNN2024.03 | 88.31 | — | — | — | — | |
| GREAD-AC2022.11 | 88.29 | — | — | — | — | |
| GCNIIsource=This paper re-test, splits=10 averaged2021.10 | 88.2696 | — | — | — | — | |
| GCN2020.05 | 88.2 | — | — | — | — | |
| WRGATtrials=102022.05 | 88.2 | — | — | — | — | |
| WRGAT2022.11 | 88.2 | — | — | — | — | |
| WRGAT2024.05 | 88.2 | — | — | — | — | |
| WRGAT2023.05 | 88.2 | — | — | — | — | |
| WRGAT2023.05 | 88.2 | — | — | — | — | |
| PhenomNN2025.10 | 88.12 | — | — | — | — | |
| GCNType=Manual GNNs2021.12 | 88.11 | — | — | — | — | |
| NLGCN2020.05 | 88.1 | — | — | — | — | |
| BLEND2022.02 | 88.09 | — | — | — | — | |
| BLEND2022.11 | 88.09 | — | — | — | — | |
| BLENDcontinuous-time=true2025.11 | 88.09 | — | — | — | — | |
| GCNIINumber of Layers=642025.01 | 88.07 | — | — | — | — | |
| FAGCN2023.05 | 88.05 | — | — | — | — | |
| ACM-GCN+2023.05 | 88.05 | — | — | — | — | |
| DSHNLight2025.10 | 88.02 | — | — | — | — | |
| GCNIIEvaluation protocol=fully-supervised2021.05 | 88.01 | — | — | — | — | |
| GCNII*Layers=642021.08 | 88.01 | — | — | — | — | |
| GCNII*source=Chen et al. (2020), Number of layers=642021.10 | 88.01 | — | — | — | — | |
| GREAD-FB2022.11 | 88.01 | — | — | — | — | |
| GREAD-FB*2022.11 | 88.01 | — | — | — | — | |
| ANS-GT2023.08 | 88 | — | — | — | — | |
| SimP-GCN2023.12 | 87.99 | — | — | — | — | |
| GGCN2022.06 | 87.95 | — | — | — | — | |
| GPRGNN2022.06 | 87.95 | — | — | — | — | |
| GPR-GNNtrials=102022.05 | 87.95 | — | — | — | — | |
| GGCNtrials=102022.05 | 87.95 | — | — | — | — | |
| GGCNLayers=Best across layers2021.02 | 87.95 | — | — | — | — | |
| GPRGNNLayers=Best across layers2021.02 | 87.95 | — | — | — | — |