Membership Inference Attack on CIFAR10
100Balanced AccuracyFedGA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FedGA2025.12 | 100 | 100 | |
| FedKD2025.12 | 66.6 | 75 | |
| Song et al.2026.01 | 65 | 64.4 | |
| MINT2026.01 | 64.9 | 72.8 | |
| Yeom et al.2026.01 | 64.7 | 64.6 | |
| Watson et al.2026.01 | 63.1 | 67.7 | |
| Curv ZO NLLshadow models=642024.07 | 61.92 | 68.82 | |
| Salem et al.2026.01 | 61 | 62.8 | |
| Ye et al.shadow models=642024.07 | 59.62 | 67.3 | |
| Carlini et al.shadow models=642024.07 | 58.23 | 61.73 | |
| FedR2S2025.12 | 57.9 | 64.8 | |
| Sablayrolles et al.shadow models=642024.07 | 56.65 | 61.5 | |
| Song and Mittalshadow models=642024.07 | 55.63 | 60.42 | |
| Yeom et al.shadow models=642024.07 | 55.57 | 60.44 | |
| FedRetrainstatus=Gold2025.12 | 55.3 | 58.3 | |
| Curv ZO LRshadow models=642024.07 | 55 | 58.89 | |
| Watson et al.shadow models=642024.07 | 54.86 | 58.58 | |
| Ye et al.2026.01 | 52.7 | 63.2 | |
| REMISVFU2025.12 | 52.5 | 57.6 |