privGAN: Protecting GANs from membership inference attacks at low cost to utility
Generative Adversarial Networks (GANs) have made releasing of synthetic images a viable approach to share data without releasing the original dataset. It has been shown that such synthetic data can be used for a variety of downstream tasks such as training classifiers that would otherwise require th...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Sciendo
2021-07-01
|
| Series: | Proceedings on Privacy Enhancing Technologies |
| Subjects: | |
| Online Access: | https://doi.org/10.2478/popets-2021-0041 |