Consensual and Privacy-Preserving Sharing of Multi-Subject and Interdependent Data

Individuals share increasing amounts of personal data online. This data often involves–or at least has privacy implications for–data subjects other than the individual who shares it (e.g., photos, genomic data) and the data is shared without their consent. A popular example, with dramatic consequences, is revenge pornography. In this paper, we propose ConsenShare, a system for sharing, in a consensual (wrt the data subjects) and privacy-preserving (wrt both service providers and other individuals) way, data involving subjects other than the uploader. We describe a complete design and implementation of ConsenShare for photos, which relies on image processing and cryptographic techniques, as well as on a two-tier architecture (one entity for detecting the data subjects and contacting them; one entity for hosting the data and for collecting consent). We benchmark the performance (CPU and bandwidth) of ConsenShare by using a dataset of 20k photos from Flickr. We also conduct a survey targeted at Facebook users (N = 321). Our results are quite encouraging: The experimental results demonstrate the feasibility of our approach (i.e., acceptable overheads) and the survey results demonstrate potential interest from the users.

Interdependent and Multi-Subject Privacy: Threats, Analysis and Protection

Alexandra-Mihaela Olteanu

In Alan Westin's generally accepted definition of privacy, he describes it as an individual's right 'to control, edit, manage, and delete information about them[selves] and decide when, how, and to what extent information is communicated to others.' Therefore, privacy is an individual and independent human right. The great Mahatma Gandhi once said that 'interdependence is and ought to be as much the ideal of man as selfsufficiency. Man is a social being.' To ensure this independent right to inherently social beings, it will be difficult, if not impossible. This is especially true as today's world is highly interconnected, technology evolves rapidly, data sharing is increasingly abundant, and regulations do not provide sufficient guidance in the realm of interdependency. In this thesis, we explore the topic of interdependent privacy from an adversarial point of view by exposing threats, as well as from an end-user point of view, by exploring awareness, preferences and privacy protection needs. First, we quantify the effect of co-locations on location privacy, considering an adversary such as a social-network operator that has access to this information: Not only can a user be localized due to her reported locations and mobility patterns, but also due to those of her friends (and the friends of her friends and so on). We formalize this problem and propose effective inference algorithms that substantially reduce the complexity of localization attacks that make use of co-locations. Our results show that an adversary can effectively incorporate co-locations in attacks to substantially reduce users' location privacy; this exposes a real and severe threat. Second, we investigate the interplay between the privacy risks and the social benefits of users when sharing (co-)locations on OSNs. We propose a game-theoretic framework for analyzing users' strategic behaviors. We conduct a survey of Facebook users and quantify their benefits of sharing vs. viewing information and their preference for privacy vs. benefits. Our survey exposes deficits in users' awareness of privacy risks in OSNs. Our results further show how users' individual preferences influence, sometimes in a negative way, each other's decisions. Third, we consider various types of interdependent and multi-subject data (photo, colocation, genome, etc.) that often have privacy implications for data subjects other than the uploader, yet can be shared without their consent or awareness. We propose a system for sharing such data in a consensual and privacy-preserving manner. We implement it in the case of photos, by relying on image-processing and cryptographic techniques, as well as on a two-tier architecture. We conduct a survey of Facebook users; it indicates that there is interest in such a system, and that users have increasing privacy concerns due to prejudice or discrimination that they have been or could still easily be exposed to. In conclusion, this thesis provides new insights on users' privacy in the context of interdependence and constitutes a step towards the design of novel privacy-protection mechanisms. It should be seen as a warning message for service providers and regulatory institutions: Unless the interdependent aspects of privacy are considered, this fundamental human right can never be guaranteed.

EPFL2019

Reactive and Proactive Sharing Across Concurrent Analytical Queries

Anastasia Ailamaki, Manoussos Gavriil Athanassoulis, Matthaios Alexandros Olma, Iraklis Psaroudakis

Today an ever increasing amount of data is collected and analyzed by researchers, businesses, and scientists in data warehouses (DW). In addition to the data size, the number of users and applications querying data grows exponentially. The increasing concurrency is itself a challenge in query execution, but also introduces an opportunity favoring synergy between concurrent queries. Traditional execution engines of DW follows a query-centric approach, where each query is optimized and executed independently. On the other hand, workloads with increased concurrency have several queries with common parts of data and work, creating the opportunity for sharing among concurrent queries. Sharing can be reactive to the inherently existing sharing opportunities, or proactive by redesigning query operators to maximize the sharing opportunities. This demonstration showcases the impact of proactive and reactive sharing by comparing and integrating representative state-of-the-art techniques: Simultaneous Pipelining (SP), for reactive sharing, which shares intermediate results of common sub-plans, and Global Query Plans (GQP) for proactive sharing, which build and evaluate a single query plan with shared operators. We visually demonstrate, in an interactive interface, the behavior of both sharing approaches on top of a state-of-the-art storage engine using the original prototypes. We show that pull-based sharing for SP eliminates the serialization point imposed by the original push-based approach. Then, we compare, through a sensitivity analysis, the performance of SP and GQP. Finally, we show that SP can improve the performance of GQP for a query mix with common sub-plans.

2014

Privacy-Preserving Federated Analytics using Multiparty Homomorphic Encryption

David Jules Froelicher

Analyzing and processing data that are siloed and dispersed among multiple distrustful stakeholders is difficult and can even become impossible when the data are sensitive or confidential. Current data-protection and privacy regulations highly restrict the sharing and outsourcing of personal information among stakeholders that are in different jurisdictions. Sharing data is, however, required in many domains. The medical sector is a paradigmatic example: Privacy is paramount and data sharing is needed in numerous applications where data is scarce and scattered among multiple stakeholders around the world. Existing privacy-preserving solutions for federated analytics (FA) rely either (1) on data centralization or outsourcing to a limited number of entities, which incur multiple security and trust issues, or (2) on the exchange of cleartext aggregated and optionally obfuscated data, which can leak personal information or introduce bias in the final result. In this thesis, our goal is (1) to propose privacy-preserving federated solutions for exploration, and for statistical and machine-learning analyses on data held by multiple distrustful stakeholders, and (2) to analyze and evaluate the proposed systems, thus showing that they provide an efficient, secure, scalable, and accurate alternative to existing solutions for FA by proving their utility in real-world state-of-the-art biomedical studies. We rely on multiparty homomorphic encryption (MHE). MHE combines secure multiparty computation (SMC) techniques with homomorphic encryption (HE) by pooling the advantages of both SMC and HE, i.e., interactivity and flexibility, and by minimizing their disadvantages, i.e., difficulty in scaling to a large number of parties and computation complexity.First, we design UnLynx, a system that enables privacy-preserving federated data exploration on a distributed dataset held by multiple data-providers (DPs), where N-1 out of N of the nodes performing the computations can be malicious. We build interactive protocols by relying on ElGamal additive homomorphic encryption (AHE) and ensure that each untrusted-node operation can be publicly verified by means of zero-knowledge proofs (ZKPs). We then explore how statistics, e.g., standard deviation and variance, can be computed by relying on AHE and ZKPs through the design of another system named Drynx. In Drynx, we also explore how to limit the influence of an entity that inputs wrong data in the system, and we propose an efficient federated solution for correctness verification.We propose Spindle a solution for secure cooperative gradient descent on federated data that we instantiate for the privacy-preserving training and oblivious evaluation of generalized linear models. Spindle covers the entire machine-learning workflow, as it enables oblivious predictions to be performed on a trained model that remains secret. It ensures both data and model confidentiality in a passive adversarial model in which N-1 out of N DPs can collude.Finally, we demonstrate that the solutions proposed in this thesis can be efficient enablers for large-scale, sensitive, multi-site biomedical studies. We design and test, by replicating recent medical studies, secure workflows for the federated execution of computations that span from analyses with low computational complexity, such as survival analyses, to analyses with high computational complexity such as genome-wide association studies on millions of variants.