Farhad Farokhi of University of Melbourne

A common, sometimes implicit, assumption within the privacy literature is the use of randomization. Differential privacy and identifiability, two popular and related definitions from the privacy literature, assume the use of randomized functions. Information-theoretic privacy often falls back on measures, such as entropy and mutual information, that are only defined for random variables. However, many popular heuristic-based privacy-preserving methods, such as k-anonymity, are deterministic. They employ deterministic mappings, such as suppression and generalization, and are applied to non-stochastic datasets. Randomized, or stochastic, privacy-preserving policies have been shown to cause problems, such as un-truthfulness and generation of unreasonable/unrealistic outputs, which can be undesirable in practice. Motivated by these observations, in this talk, we develop deterministic privacy frameworks. We first borrow some useful concepts from the sparse literature on non-stochastic information theory and generalize other concepts from information theory and signal processing literatures to the non-stochastic setting. Particularly, we use a worst-case measure of uncertainty and the maximin information to measure private information leakage. We pose the problem of determining deterministic privacy-preserving policies as maximization of the measure of privacy subject to a constraint on the worst-case distortion. This way, we can show that the optimal privacy-preserving policy is a piecewise constant function in the form of a quantization operator. We show that we can use our privacy metrics to analyse k-anonymity, proving that it is in fact not privacy-preserving. Finally, we take a slight detour to develop a theory of non-stochastic hypothesis testing for generalizing the concept of identifiability from the privacy literature to the non-scholastic setting.  This talk is mainly based on two papers: a journal paper in IEEE Transactions on Information Forensics and Security (https://arxiv.org/abs/1810.11153) and a forthcoming conference paper in CDC (https://arxiv.org/abs/1904.07377).

Biography:

Farhad Farokhi is a Research Scientist at the Information Security and Privacy Group at CSIRO's Data61 and a Research Fellow at the Department of Electrical and Electronic Engineering at the University of Melbourne. In 2014, he received his PhD degree in Automatic Control from KTH Royal Institute of Technology, Sweden. During his PhD studies, he was a visiting researcher at the University of California at Berkeley and the University of Illinois at Urbana-Champaign. Farhad has been the recipient of the VESKI Victoria Fellowship from the Victorian State Government as well as the McKenzie Fellow and the 2015 Early Career Researcher Award from the University of Melbourne. He was a finalist in the 2014 European Embedded Control Institute (EECI) PhD Award. He has been part of numerous projects on data privacy and cyber-security funded by the Defence Science and Technology Group (DSTG), the Department of the Prime Minister and Cabinet (PMC), the Department of Environment and Energy (DEE), and CSIRO in Australia. His research interests include security and privacy in cyber-physical systems, such as smart grids and intelligent transportation systems.