Node Classification on Cora-ML
97.63AccuracyIMPRESS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| IMPRESSWays=2, Shots=52026.04 | 97.63 | — | |
| IMPRESSWays=2, Shots=32026.04 | 97.1 | — | |
| VISIONn_way=2, k_shot=52026.05 | 93.71 | — | |
| NaQWays=2, Shots=52026.04 | 92.56 | — | |
| G-MetaWays=2, Shots=52026.04 | 92.16 | — | |
| G-Metan_way=2, k_shot=52026.05 | 92.16 | — | |
| Meta-GPSWays=2, Shots=52026.04 | 91.66 | — | |
| Meta-GPSn_way=2, k_shot=52026.05 | 91.66 | — | |
| VISIONn_way=2, k_shot=32026.05 | 91.6 | — | |
| X-FNCWays=2, Shots=52026.04 | 90.36 | — | |
| X-FNCn_way=2, k_shot=52026.05 | 90.36 | — | |
| GraphPromptn_way=2, k_shot=52026.05 | 90.32 | — | |
| VNTWays=2, Shots=52026.04 | 89.26 | — | |
| VNTn_way=2, k_shot=52026.05 | 89.26 | — | |
| SMILEn_way=2, k_shot=52026.05 | 89.22 | — | |
| G-MetaWays=2, Shots=32026.04 | 88.68 | — | |
| G-Metan_way=2, k_shot=32026.05 | 88.68 | — | |
| Meta-GPSWays=2, Shots=32026.04 | 87.91 | — | |
| Meta-GPSn_way=2, k_shot=32026.05 | 87.91 | — | |
| GraphPromptn_way=2, k_shot=32026.05 | 87.52 | — | |
| NaQWays=2, Shots=32026.04 | 87.37 | — | |
| GNN2025.05 | 87.06 | — | |
| SSN2025.05 | 86.64 | — | |
| Dir-GNN2025.05 | 86.6 | — | |
| MPSNN2025.05 | 86.58 | — | |
| APPNPAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 86.3 | — | |
| VNTWays=2, Shots=32026.04 | 86.22 | — | |
| VNTn_way=2, k_shot=32026.05 | 86.22 | — | |
| GRANDBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 86.1 | — | |
| GRANDAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 86.1 | — | |
| AdaGCN2019.08 | 85.97 | — | |
| RUNG-l1Attack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 85.8 | — | |
| SMILEn_way=2, k_shot=32026.05 | 85.72 | — | |
| RGCNBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 85.7 | — | |
| RGCNAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 85.7 | — | |
| X-FNCWays=2, Shots=32026.04 | 85.62 | — | |
| X-FNCn_way=2, k_shot=32026.05 | 85.62 | — | |
| ProGNNBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 85.6 | — | |
| ProGNNAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 85.6 | — | |
| VISIONn_way=2, k_shot=12026.05 | 85.44 | — | |
| TLPWays=2, Shots=32026.04 | 85.32 | — | |
| TLPn_way=2, k_shot=32026.05 | 85.32 | — | |
| PPNPImplementation=Literature2019.08 | 85.29 | — | |
| Meta-BPn_way=2, k_shot=52026.05 | 85.19 | — | |
| APPNPImplementation=Literature2019.08 | 85.09 | — | |
| GCNNumber of runs=1002022.04 | 85 | — | |
| GCNBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 85 | — | |
| SoftMedianBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 85 | — | |
| GCNAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 85 | — | |
| SoftMedianAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 85 | — | |
| TRI-GNNDfiltration function=node degree2021.10 | 84.98 | 0.49 | |
| TRI-GNNAfiltration function=node attributes2021.10 | 84.8 | 0.55 | |
| EvenNetAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 84.8 | — | |
| SuperGATMxNumber of runs=1002022.04 | 84.7 | — | |
| Pro-GATBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 84.6 | — | |
| RUNGAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 84.6 | — | |
| SuperGATSDNumber of runs=1002022.04 | 84.5 | — | |
| HANGAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 84.5 | — | |
| PPNPImplementation=Ours2019.08 | 84.39 | — | |
| GAT2019.08 | 84.37 | — | |
| LFGCN2021.10 | 84.35 | 0.57 | |
| APPNPImplementation=Ours2019.08 | 84.28 | — | |
| TWIRLSAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 84.2 | — | |
| GARNETAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 84 | — | |
| NoisyGNNAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 83.9 | — | |
| GMNN2021.10 | 83.72 | 0.9 | |
| GraphSAGENumber of runs=1002022.04 | 83.7 | — | |
| Jaccard-GCNBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 83.7 | — | |
| Jaccard-GCNAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 83.7 | — | |
| GraphCONAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 83.7 | — | |
| LGCNS2021.10 | 83.63 | 0.55 | |
| GATBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 83.5 | — | |
| GATAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 83.5 | — | |
| GCNOptimization=Early stopping2019.08 | 83.41 | — | |
| SPAGAN2021.10 | 83.31 | 0.53 | |
| APPNP2021.10 | 83.31 | 0.53 | |
| GPNWays=2, Shots=52026.04 | 83.21 | — | |
| GPNn_way=2, k_shot=52026.05 | 83.21 | — | |
| GATNumber of runs=1002022.04 | 83.2 | — | |
| GAT2021.10 | 83.11 | 0.7 | |
| GNNGuardBudget=0%, Attack=PGD adaptive attack, Number of random splits=52024.10 | 83.1 | — | |
| GNNGuardAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 83.1 | — | |
| ARMA2021.10 | 82.8 | 0.63 | |
| RGCN2021.10 | 82.8 | 0.6 | |
| TWIRLS-TAttack Budget=0% (Clean), Attack Type=Adaptive global attack2023.11 | 82.8 | — | |
| PEGN-RC2021.10 | 82.7 | 0.5 | |
| JK2019.08 | 82.69 | — | |
| V.GCNDescription=Vanilla GCN2019.08 | 82.3 | — | |
| ScaleNetType=Ours, loop_α, β, γ=1_2,-1,-12024.11 | 82.3 | — | |
| N-GCN2019.08 | 82.25 | — | |
| 1iGu2Type=Symmetric (Ours)2024.11 | 82.1 | — | |
| 1iGType=Symmetric (Ours)2024.11 | 82 | — | |
| MixHop2021.10 | 81.9 | 0.81 | |
| SymType=Symmetric2024.11 | 81.9 | — | |
| VPN2021.10 | 81.89 | 0.57 | |
| APPNPType=Base models2024.11 | 81.8 | — | |
| SAGEType=Base models2024.11 | 81.7 | — | |
| 1iGi2Type=Symmetric (Ours)2024.11 | 81.7 | — | |
| GCN2021.10 | 81.5 | 0.4 | |
| ChebNet2021.10 | 81.45 | 1.21 |