Towards Statistical Queries over Distributed Private User Data

By Ruichuan Chen, Alexey Reznichenko, Paul Francis, Johannes Gehrke. 2012

130 citations recorded by the source corpus.

Subfields: Federated & Distributed Learning, Privacy & Security

How researchers see this paper

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

In Federated & Distributed Learning

In Privacy & Security

Citations in the corpus

Honeycrisp: large-scale differentially private aggregation without a trusted core (2019, background): “There is now a substantial literature on differential privacy, both on the theoretical foundations [28] and on practical implementations [20, 55, 62, 68].”
Prochlo: Strong Privacy for Analytics in the Crowd (2017, background): “Other cryptography-based privacy-protection systems, such as PDDP, Prio, and Secure Aggregation [11, 15, 18], mostly share ESA’s goals but differ greatly in their approach.”
Prio: Private, Robust, and Scalable Computation of Aggregate Statistics (2017, background): “A variant of the same principle is to have a trusted proxy (as in SuLQ [21] and PDDP [34]) or a set of minimally trusted servers [92] add noise to already-collected data.”
RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response (2014, background): “[8] and [2] implement a diﬀerentially private protocol over distributed user data by relying on an honest-but-curious proxy or data aggregator bound by certain commitments.”
Efficient Private Statistics with Succinct Sketches (2015, background): “A combination of homomorphic encryption and differential privacy has been explored by Chen et al. [19], allowing third parties to gather web analytics.”
Verifiable differential privacy (2015, background): “A few recent systems have considered richer threat models: for instance, there are some solutions [14, 33, 36] that can process queries over multiple databases whose curators do not trust each other, while others [10] can prevent malicious subjects from providing data that would distort the results.”
DStress: Efficient Differentially Private Computations on Distributed Data (2017, background): “Later systems added support for distributed data, but they typically focus on a specific class of queries: for instance, PDDP [16] can build histograms, and DJoin [51] can process certain types of joins.”
UnLynx: A Decentralized System for Privacy-Conscious Data Sharing (2017, background): “[17] propose a solution with multiple parties that work together with a trusted proxy in order to add a distributively created noise.”
Non-tracking web analytics (2012, background): “proposed a proxy-based system (PDDP) for achieving differential privacy in a distributed environment [32].”
SplitX: high-performance private analytics (2013, methodology): “SplitX operates under a similar trust model as several recent analytics proposals [4, 9, 20], i.”
Private spatial data aggregation in the local setting (2016, background): “Chen et al. [17] develop a distributed differentially private system to support statistical queries over multiple clients’ data via an honest-but-curious proxy.”
DJoin: Differentially Private Join Queries over Distributed Databases (2012, unknown): “Several differentially private query processors, including PINQ [23], Airavat [32], Fuzz [16], and PDDP [6], have been developed and are available today.”
Dynamic Differential Privacy for ADMM-Based Distributed Classification Learning (2017, background): “Such works include multi-party distributed data aggregation [14] and private multi-party computations [15].”
What Does The Crowd Say About You? Evaluating Aggregation-based Location Privacy (2017, background): “ herently unique to the user, and the identities of the subjects generating the traces can often be recovered [11,20,52]. In some cases, mobility models can be trained using only aggregate statistics [10,31,37], e.g., how many people are in a certain location at a given time. Therefore, a common app”
PROCHLO: Strong Privacy for Analytics in the Crowd (2017, background): “Other cryptography-based privacy-protection systems, such as PDDP, Prio, and Secure Aggregation [11, 15, 18], mostly share ESA’s goals but differ greatly in their approach.”
Differential Privacy by Typing in Security Protocols (2013, background): “aggregation algorithms [9]–[13], and local learning algorithms [14]) and on computational definitions of differential privacy against polynomially-bounded opponents [8].”
Stadium: A Distributed Metadata-Private Messaging System (2017, background): “Another related line of work is user privacy for untrusted aggregators wishing to learn summary statistics over user populations [21, 11].”
Distributed Differential Privacy and Applications (2015, methodology): “(2010)—or that distributed queries can be broken into several subqueries that can each be answered using only one of the databases (Dwork et al., 2006a; Rastogi and Nath, 2010; Chen et al., 2012; Götz and Nath, 2011).”
Differentially private data aggregation with optimal utility (2014, background): “The former rely on trusted or honest-but-curious (HbC) 1 servers [6, 15, 32] to compute noise perturbation.”