Daniel Bernau

Abstract

We present two information leakage attacks that outperform previous work on membership inference against generative models. The first attack allows membership inference without assumptions on the type of the generative model. Contrary to previous evaluation metrics for generative models, like Kernel Density Estimation, it only considers samples of the model which are close to training data records. The second attack specifically targets Variational Autoencoders, achieving high membership inference accuracy. Furthermore, previous work mostly considers membership inference adversaries who perform single record membership inference. We argue for considering regulatory actors who perform set membership inference to identify the use of specific datasets for training. The attacks are evaluated on two generative model architectures, Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs), trained on standard image datasets. Our results show that the two attacks yield success rates superior to previous work on most data sets while at the same time having only very mild assumptions. We envision the two attacks in combination with the membership inference attack type formalization as especially useful. For example, to enforce data privacy standards and automatically assessing model quality in machine learning as a service setups. In practice, our work motivates the use of GANs since they prove less vulnerable against information leakage attacks while producing detailed samples.

Abstract

We present two information leakage attacks that outperform previous work on membership inference against generative models. The first attack allows membership inference without assumptions on the type of the generative model. Contrary to previous evaluation metrics for generative models, like Kernel Density Estimation, it only considers samples of the model which are close to training data records. The second attack specifically targets Variational Autoencoders, achieving high membership inference accuracy. Furthermore, previous work mostly considers membership inference adversaries who perform single record membership inference. We argue for considering regulatory actors who perform set membership inference to identify the use of specific datasets for training. The attacks are evaluated on two generative model architectures, Generative Adversarial Networks (GANs) and Variational Autoen-coders (VAEs), trained on standard image datasets. Our results show that the two attacks yield success rates superior to previous work on most data sets while at the same time having only very mild assumptions. We envision the two attacks in combination with the membership inference attack type formalization as especially useful. For example, to enforce data privacy standards and automatically assessing model quality in machine learning as a service setups. In practice, our work motivates the use of GANs since they prove less vulnerable against information leakage attacks while producing detailed samples.

Abstract

There has been a large number of contributions on privacy-preserving smart metering with Differential Privacy, addressing questions from actual enforcement at the smart meter to billing at the energy provider. However, exploitation is mostly limited to application of cryptographic security means between smart meters and energy providers. We illustrate along the use case of privacy preserving load forecasting that Differential Privacy is indeed a valuable addition that unlocks novel information flows for optimization. We show that (i) there are large differences in utility along three selected forecasting methods, (ii) energy providers can enjoy good utility especially under the linear regression benchmark model, and (iii) households can participate in privacy preserving load forecasting with an individual re-identification risk < 60%, only 10% over random guessing.

Abstract

In cyber-physical systems sensors data should be anonymized at the source. Local data perturbation with differential privacy guarantees can be used, but the resulting utility is often (too) low. In this paper we contribute an algorithm that combines local, differentially private data perturbation of sensor streams with highly accurate outlier detection. We evaluate our algorithm on synthetic data. In our experiments we obtain an accuracy of 80% with a differential privacy value of ��=0.1for well separated outliers.

Abstract

Object-level tracking along supply chains, enabled by the low-cost and wide availability of Radio Frequency Identification (RFID) technology, permits companies to collect large amounts of data (e.g., time, location, handling) about the goods they produce. Combining the data collected by the different companies along a supply chain can provide considerable advantages to all of them. However, such a sharing sometimes needs to be selective. Indeed, companies may need to keep some information about their business operations secret. In this paper, we propose a solution that enables selective sharing of data collected along a supply chain. Our solution uses the services offered by cloud providers for sharing data among companies, and relies on selective encryption for enforcing access restrictions over such data.