Membership Inference Advantage Bounding on Uniform prior size 100
1.2Advantage Upper BoundReRo bound
Evaluation results
| Method | Links | |
|---|---|---|
| ReRo boundsigma=32023.02 | 1.2 | |
| ReRo boundsigma=2.52023.02 | 1.8 | |
| ReRo boundsigma=22023.02 | 2.4 | |
| ReRo boundsigma=1.52023.02 | 3.5 | |
| Guo et al. (2022b) (MC)sigma=3, Monte-Carlo approximation=true2023.02 | 4.9 | |
| Guo et al. (2022b) (MC)sigma=2.5, Monte-Carlo approximation=true2023.02 | 6.2 | |
| Guo et al. (2022b)sigma=32023.02 | 7.6 | |
| ReRo boundsigma=12023.02 | 7.7 | |
| Guo et al. (2022b) (MC)sigma=2, Monte-Carlo approximation=true2023.02 | 8.1 | |
| Guo et al. (2022b)sigma=2.52023.02 | 9.7 | |
| Guo et al. (2022b) (MC)sigma=1.5, Monte-Carlo approximation=true2023.02 | 12 | |
| Guo et al. (2022b)sigma=22023.02 | 13.1 | |
| Guo et al. (2022b)sigma=1.52023.02 | 19.5 | |
| Guo et al. (2022b) (MC)sigma=1, Monte-Carlo approximation=true2023.02 | 21 | |
| Guo et al. (2022b)sigma=12023.02 | 34.6 | |
| ReRo boundsigma=0.52023.02 | 36.2 | |
| Guo et al. (2022b) (MC)sigma=0.5, Monte-Carlo approximation=true2023.02 | 54.9 | |
| Guo et al. (2022b)sigma=0.52023.02 | 86.1 |