Node Classification on Squirrel (average over 5 different splits)
33.64AccuracyFAGCN
Evaluation Results
| Method | Links | |
|---|---|---|
| FAGCNAttack Method=MinMax, Perturb Ratio=20%, Surrogate Model=GCN, Weight Decay=0, Number of Layers=2, epsilon={0.3, 0.4, 0.5}2022.05 | 33.64 | |
| EvenNetAttack Method=MinMax, Perturb Ratio=20%, Surrogate Model=GCN, Weight Decay=0, alpha=0.1, K=102022.05 | 33.21 | |
| GPRGNNAttack Method=MinMax, Perturb Ratio=20%, Surrogate Model=GCN, Weight Decay=0, alpha=0.1, K=102022.05 | 32.47 | |
| GCNAttack Method=MinMax, Perturb Ratio=20%, Surrogate Model=GCN, Weight Decay=02022.05 | 31.16 | |
| MLPAttack Method=MinMax, Perturb Ratio=20%, Surrogate Model=GCN, Weight Decay=02022.05 | 30.31 | |
| H2GCNAttack Method=MinMax, Perturb Ratio=20%, Surrogate Model=GCN, Weight Decay=0, Number of Layers=22022.05 | 28.41 | |
| EGCNGuardAttack Method=MinMax, Perturb Ratio=20%, Surrogate Model=GCN, Weight Decay=0, Threshold=0.12022.05 | 27.34 |