TableGAN-MCA: Evaluating Membership Collisions of GAN-Synthesized Tabular Data Releasing

Hu, Aoting; Xie, Renjie; Lü, Zhigang; Hu, Aiqun; Xue, Minhui

doi:10.1145/3460120.3485251

Cited by 13 publications

(7 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Membership Inference Attack. Recent studies show that GANs are vulnerable to Membership Inference Attacks (MIAs) [3,21,22], which aim to identify if a given data record was used to train the GANs. Hu et al [22] studies Membership Collisions Attack (MCA) on tabular GANs.…”

Section: Security Analysismentioning

confidence: 99%

“…Recent studies show that GANs are vulnerable to Membership Inference Attacks (MIAs) [3,21,22], which aim to identify if a given data record was used to train the GANs. Hu et al [22] studies Membership Collisions Attack (MCA) on tabular GANs. They assume there is an intersection between the synthetic dataset and the real training dataset, and the proposed attack model successfully recovers some training data from the published synthetic dataset.…”

Section: Security Analysismentioning

confidence: 99%

See 1 more Smart Citation

GTV: Generating Tabular Data via Vertical Federated Learning

Zhang¹,

Han²,

Moorsel³

et al. 2023

Preprint

View full text Add to dashboard Cite

Generative Adversarial Networks (GANs) have achieved state-ofthe-art results in tabular data synthesis, under the presumption of direct accessible training data. Vertical Federated Learning (VFL) is a paradigm which allows to distributedly train machine learning model with clients possessing unique features pertaining to the same individuals, where the tabular data learning is the primary use case. However, it is unknown if tabular GANs can be learned in VFL. Demand for secure data transfer among clients and GAN during training and data synthesizing poses extra challenge. Conditional vector for tabular GANs is a valuable tool to control specific features of generated data. But it contains sensitive information from real data -risking privacy guarantees. In this paper, we propose GTV, a VFL framework for tabular GANs, whose key components are generator, discriminator and the conditional vector. GTV proposes an unique distributed training architecture for generator and discriminator to access training data in a privacy-preserving manner. To accommodate conditional vector into training without privacy leakage, GTV designs a mechanism training-with-shuffling to ensure that no party can reconstruct training data with conditional vector. We evaluate the effectiveness of GTV in terms of synthetic data quality, and overall training scalability. Results show that GTV can consistently generate high-fidelity synthetic tabular data of comparable quality to that generated by centralized GAN algorithm. The difference on machine learning utility can be as low as to 2.7%, even under extremely imbalanced data distributions across clients and different number of clients.

show abstract

Section: Security Analysismentioning

confidence: 99%

Section: Security Analysismentioning

confidence: 99%

GTV: Generating Tabular Data via Vertical Federated Learning

Zhang¹,

Han²,

Moorsel³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…We augment synthetic datasets with GAN [10] to better train shadow models without additional queries, which is also an extension to the application scenario of membership inference attacks in Section II. Compared with traditional neural network-based data augmentation methods such as VAE [29] and PixelRNN [30], GAN [31] has the advantages of higher quality synthetic data, more similar to the original data, and can be synthesized in large quantities in a short period of time, which is exactly what is needed for training shadow models.…”

Section: Data Augmented Shadow Model Trainingmentioning

confidence: 99%

Black-Box Based Limited Query Membership Inference Attack

et al. 2022

View full text Add to dashboard Cite

Conventional membership inference attacks usually require a large number of queries of the target model when training shadow models, and this task becomes extremely difficult when the number of queries is limited. Aiming at the problem of insufficient training data for shadow models due to the limited number of queries, we propose a membership inference attack method based on generative adversarial networks (GAN). First, we use generative adversarial networks to augment the samples obtained by a small number of queries to expand the training data of the model; Secondly, we use the improved CNN to obtain shadow models that have a higher degree of fitting on different target model structures; Finally, we evaluate the accuracy of the proposed algorithm on XgBoost, Logistic, and neural network models using public datasets MNIST and CIFAR10 in a black-box setting, and the results show that our model has an average attack accuracy of 62% and 83%, respectively. It can be seen that, compared with the existing research methods, our model can obtain better attack effects under the condition of significantly reducing the number of queries, which shows the feasibility of our proposed method in membership inference attacks.INDEX TERMS membership inference attack; generative adversarial network; black-box attack; information leak

show abstract

“…GANs have shown great success in generating high-quality synthetic images [34][35][36] indistinguishable from real images. This has encouraged the use of GANs for synthetic data generation in broader contexts, in particular in high-energy physics, where in some instances the data generation can be a computational intensive task [37][38][39][40].…”

Section: Introductionmentioning

confidence: 99%

“…This has encouraged the use of GANs for synthetic data generation in broader contexts, in particular in high-energy physics, where in some instances the data generation can be a computational intensive task [37][38][39][40]. In this regard, while GANs were developed for image generation [33], there have been attempts to adapt this approach for other formats, such as tabular data [34], time series [41], video content augmentation [42] and audio synthesis [43][44][45].…”

Section: Introductionmentioning

confidence: 99%

Generating gravitational waveform libraries of exotic compact binaries with deep learning

Freitas¹,

Herdeiro²,

Morais³

et al. 2022

Preprint

View full text Add to dashboard Cite

Current gravitational wave (GW) detections rely on the existence of libraries of theoretical waveforms. Consequently, finding new physics with GWs requires libraries of non-standard models, which are computationally demanding. We discuss how deep learning frameworks can be used to generate new waveforms "learned" from a simulation dataset obtained, say, from numerical relativity simulations. Concretely, we use the WaveGAN architecture of a generative adversarial network (GAN). As a proof of concept we provide this neural network (NN) with a sample of (> 500) waveforms from the collisions of exotic compact objects (Proca stars), obtained from numerical relativity simulations. Dividing the sample into a training and a validation set, we show that after a sufficiently large number of training epochs the NN can produce from 12% to 25% of the synthetic waveforms with an overlapping match of at least 95% with the ones from the validation set. We also demonstrate that a NN can be used to predict the overlapping match score, with 90% of accuracy, of new synthetic samples. These are encouraging results for using GANs for data augmentation and interpolation in the context of GWs, to cover the full parameter space of, say, exotic compact binaries, without the need of intensive numerical relativity simulations.

show abstract

TableGAN-MCA: Evaluating Membership Collisions of GAN-Synthesized Tabular Data Releasing

Cited by 13 publications

References 25 publications

GTV: Generating Tabular Data via Vertical Federated Learning

GTV: Generating Tabular Data via Vertical Federated Learning

Black-Box Based Limited Query Membership Inference Attack

Generating gravitational waveform libraries of exotic compact binaries with deep learning

Contact Info

Product

Resources

About