Computational Privacy
Lead Research Organisation:
Imperial College London
Department Name: Computing
Abstract
Computational Privacy
Organisations
Publications
Chatalic A.
(2021)
Compressive Learning with Privacy Guarantees
in Information and Inference
Gadotti A
(2019)
When the Signal is in the Noise: Exploiting Diffix's Sticky Noise
Schellekens V
(2019)
Differentially Private Compressive K-means
Studentship Projects
Project Reference | Relationship | Related To | Start | End | Student Name |
---|---|---|---|---|---|
EP/N509486/1 | 30/09/2016 | 30/03/2022 | |||
1964895 | Studentship | EP/N509486/1 | 30/09/2017 | 30/03/2021 | Florimond Houssiau |
Description | This project aims at evaluating privacy risks of modern environments. Specifically, this has been split in two research questions: (1) Can users protect the privacy of their Web search queries through "obfuscation" (hiding them among artificially generated queries)? (2) How vulnerable are people in highly-connected environments to attackers compromising users of the network? In (1), we developed a theoretical model to represent obfuscation, and found a mathematical bound on the privacy guarantees that such techniques can give. In (2), we developed a framework for node-intrusion attacks, and used it to model recent attacks (such as the Cambridge Analytica story). |
Exploitation Route | The work on query obfuscation can be used to evaluate qualitatively the privacy guarantees of obfuscators, which currently rely on ad hoc metrics. It can also help guide the design of better mechanisms. The framework we develop for privacy in networked environments might help to guide privacy regulation, and establish a robust way to quantify privacy risks on social networks or of surveillance regulation. It can also be extended to larger classes of graphes, by exploring the relation of privacy risk with graph structure. |
Sectors | Digital/Communication/Information Technologies (including Software) Government Democracy and Justice |
URL | https://cpg.doc.ic.ac.uk/blog/cambridge-analytica-is-only-the-beginning/ |