DP-CGAN: Differentially Private Synthetic Data and Label Generation

Torkzadehmahani, Reihaneh; Kairouz, Peter; Paten, Benedict

doi:10.1109/cvprw.2019.00018

Cited by 141 publications

(125 citation statements)

References 42 publications

(83 reference statements)

Supporting

Mentioning

113

Contrasting

Order By: Relevance

“…Reliable measurements need to be developed in the future to assure complete anonymity of the source individuals given the released AGs. In particular, we will investigate whether the differential privacy framework is performant in the context of large population genomics datasets [50,51].…”

Section: Discussionmentioning

confidence: 99%

Creating artificial human genomes using generative neural networks

et al. 2021

View full text Add to dashboard Cite

Generative models have shown breakthroughs in a wide spectrum of domains due to recent advancements in machine learning algorithms and increased computational power. Despite these impressive achievements, the ability of generative models to create realistic synthetic data is still under-exploited in genetics and absent from population genetics. Yet a known limitation in the field is the reduced access to many genetic databases due to concerns about violations of individual privacy, although they would provide a rich resource for data mining and integration towards advancing genetic studies. In this study, we demonstrated that deep generative adversarial networks (GANs) and restricted Boltzmann machines (RBMs) can be trained to learn the complex distributions of real genomic datasets and generate novel high-quality artificial genomes (AGs) with none to little privacy loss. We show that our generated AGs replicate characteristics of the source dataset such as allele frequencies, linkage disequilibrium, pairwise haplotype distances and population structure. Moreover, they can also inherit complex features such as signals of selection. To illustrate the promising outcomes of our method, we showed that imputation quality for low frequency alleles can be improved by data augmentation to reference panels with AGs and that the RBM latent space provides a relevant encoding of the data, hence allowing further exploration of the reference dataset and features for solving supervised tasks. Generative models and AGs have the potential to become valuable assets in genetic studies by providing a rich yet compact representation of existing genomes and high-quality, easy-access and anonymous alternatives for private databases.

show abstract

Section: Discussionmentioning

confidence: 99%

Creating artificial human genomes using generative neural networks

et al. 2021

View full text Add to dashboard Cite

show abstract

“…However, the model has to sacrifice a lot in terms of sample quality and diversity of synthetic images, making them not particularly useful for practical applications. For the specific task of generating class specific images, [23] introduces a differentially private extension to conditional GAN [24]. This can improve the downstream utility for a limited set of classification tasks.…”

Section: Private Gansmentioning

confidence: 99%

privGAN: Protecting GANs from membership inference attacks at low cost to utility

Mukherjee

Trivedi

et al. 2021

Proceedings on Privacy Enhancing Technologies

View full text Add to dashboard Cite

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 the original dataset to be shared. However, recent work has shown that the GAN models and their synthetically generated data can be used to infer the training set membership by an adversary who has access to the entire dataset and some auxiliary information. Current approaches to mitigate this problem (such as DPGAN [1]) lead to dramatically poorer generated sample quality than the original non–private GANs. Here we develop a new GAN architecture (privGAN), where the generator is trained not only to cheat the discriminator but also to defend membership inference attacks. The new mechanism is shown to empirically provide protection against this mode of attack while leading to negligible loss in downstream performances. In addition, our algorithm has been shown to explicitly prevent memorization of the training set, which explains why our protection is so effective. The main contributions of this paper are: i) we propose a novel GAN architecture that can generate synthetic data in a privacy preserving manner with minimal hyperparameter tuning and architecture selection, ii) we provide a theoretical understanding of the optimal solution of the privGAN loss function, iii) we empirically demonstrate the effectiveness of our model against several white and black–box attacks on several benchmark datasets, iv) we empirically demonstrate on three common benchmark datasets that synthetic images generated by privGAN lead to negligible loss in downstream performance when compared against non– private GANs. While we have focused on benchmarking privGAN exclusively on image datasets, the architecture of privGAN is not exclusive to image datasets and can be easily extended to other types of datasets. Repository link: https://github.com/microsoft/privGAN.

show abstract

“…Our results in Theorem 4.1 theoretically show that the attack is possible in general. During the learning of the adversarial discriminator, injecting predefined noise is known to be effective to defend such attacks [41]. Meanwhile, users could quit or frequently opt-out the federated communication when the privacy budget (quantified by noise and Differential Privacy metric [7]) is low.…”

Section: Privacy Risks From Malicious Fade Usersmentioning

confidence: 99%

Federated Adversarial Debiasing for Fair and Transferable Representations

Hong

Zhu

et al. 2021

Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery &Amp; Data Mining

View full text Add to dashboard Cite

Federated learning is a distributed learning framework that is communication efficient and provides protection over participating users' raw training data. One outstanding challenge of federate learning comes from the users' heterogeneity, and learning from such data may yield biased and unfair models for minority groups. While adversarial learning is commonly used in centralized learning for mitigating bias, there are significant barriers when extending it to the federated framework. In this work, we study these barriers and address them by proposing a novel approach Federated Adversarial DEbiasing (FADE). FADE does not require users' sensitive group information for debiasing and offers users the freedom to optout from the adversarial component when privacy or computational costs become a concern. We show that ideally, FADE can attain the same global optimality as the one by the centralized algorithm. We then analyze when its convergence may fail in practice and propose a simple yet effective method to address the problem. Finally, we demonstrate the effectiveness of the proposed framework through extensive empirical studies, including the problem settings of unsupervised domain adaptation and fair learning. Our codes and pretrained models are available at: https://github.com/illidanlab/FADE.

show abstract

DP-CGAN: Differentially Private Synthetic Data and Label Generation

Cited by 141 publications

References 42 publications

Creating artificial human genomes using generative neural networks

Creating artificial human genomes using generative neural networks

privGAN: Protecting GANs from membership inference attacks at low cost to utility

Federated Adversarial Debiasing for Fair and Transferable Representations

Contact Info

Product

Resources

About