Context-Aware Generative Adversarial Privacy

By Chong Huang, Peter Kairouz, Xiao Chen, Lalitha Sankar, Ram Rajagopal. 2017

167 citations recorded by the source corpus.

Preserving the utility of published datasets while simultaneously providing provable privacy guarantees is a well-known challenge. On the one hand, context-free privacy solutions, such as differential privacy, provide strong privacy guarantees, but often lead to a significant reduction in utility. On the other hand, context-aware privacy solutions, such as information theoretic privacy, achieve an improved privacy-utility tradeoff, but assume that the data holder has access to dataset statistics. We circumvent these limitations by introducing a novel context-aware privacy framework called generative adversarial privacy (GAP). GAP leverages recent advancements in generative adversarial networks (GANs) to allow the data holder to learn privatization schemes from the dataset itself. Under GAP, learning the privacy mechanism is formulated as a constrained minimax game between two players: a privatizer that sanitizes the dataset in a way that limits the risk of inference attacks on the individuals' private variables, and an adversary that tries to infer the private variables from the sanitized dataset. To evaluate GAP's performance, we investigate two simple (yet canonical) statistical dataset models: (a) the binary data model, and (b) the binary Gaussian mixture model. For both models, we derive game-theoretically optimal minimax privacy mechanisms, and show that the privacy mechanisms learned from data (in a generative adversarial fashion) match the theoretically optimal ones. This demonstrates that our framework can be easily applied in practice, even in the absence of dataset statistics.

Subfields: Artificial Intelligence, Safety Research

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Safety Research

Citations in the corpus

Privacy in Deep Learning: A Survey (2020, background): “There is a body of work studying information-theoretic based methods for both privacy and fairness, where privacy and fairness are provided through information degradation, through obfuscation or adversarial learning and demonstrated by mutual information reduction [67, 68, 48, 69, 70, 71, 72, 73, 7”
Olympus: Sensor Privacy through Utility Aware Obfuscation (2018, unknown): “The approach proposed in [22] is not appli-”
Mobile sensor data anonymization (2018, methodology): “These techniques have been recently applied to calculate mutual information for solving optimization problems [9, 13, 23, 32].”
Adversarial representation learning for synthetic replacement of private attributes (2020, methodology): “Like in the current work, Huang et al. (2017, 2018) use adversarial learning to minimize the mutual information between the private attribute and the censored image under a distortion constraint.”
Generative Adversarial Privacy: A Data-Driven Approach to Information-Theoretic Privacy (2018, background): “Context-aware privacy notions fall under the domain of information theoretic (IT) privacy; see, for example, [10]–[15].”
Generating Differentially Private Datasets Using GANs (2018, methodology): “nts (seeds) into synthetic data points, which are then ﬁltered by a privacy test based on a plausible deniability criterion. This procedure would be rather expensive for complex data, such as images. Huang et al. (2017) introduce the notion of generative adversarial privacy and use GANs to obfuscate”
On the Robustness of Information-Theoretic Privacy Measures and Mechanisms (2018, methodology): “, the privacy mechanisms implemented using neural networks in [26].”
Privacy and Utility Preserving Sensor-Data Transformations (2019, background): “…or frequency of a common variable among users [10], with multi-dimensional data released sequentially a more practical privacy model is inferential privacy [11, 12, 13] that measures the diﬀerence between an adversary’s belief about sensitive inferences before and after observing the released data.”
Obfuscation via Information Density Estimation (2019, background): “…the image) by incorporating either complete (cf. information-theoretic privacy (Calmon and Fawaz, 2012; Asoodeh et al., 2018; Issa et al., 2018; Hsu et al., 2018; Diaz et al., 2018)) or partial (cf. generative adversarial privacy (Huang et al., 2017)) knowledge of the underlying data distribution.”