Membership Inference Attack on MAESTRO v3
0.91AUCλ-resolved Membership Inference Attack
Evaluation results
| Method | Links | ||
|---|---|---|---|
| λ-resolved Membership Inference AttackClassifier=MLP (2 hidden layers, 64-32 units), K (noise realisations)=1002026.06 | 0.91 | 56.7 | |
| PIARF2026.06 | 0.83 | 36.5 | |
| SECMIRF2026.06 | 0.72 | 13.9 | |
| NAIVERF2026.06 | 0.67 | 14.1 |