For privacy protection, the responses to the queries are systematically corrupted with an additive random noise. Using the Cramer-Rao bound, we can relate the variance of any estimator of the private database entries to the inverse of the trace of the Fisher information matrix, motivating its use as a measure of privacy. We can compute the probability density that minimizes the trace of the Fisher information (as a proxy for maximizing the measure of privacy) to find the optimal privacy-preserving policy.
- F. Farokhi, H. Sandberg,“Fisher Information as a Measure of Privacy: Preserving Privacy of Households with Smart Meters Using Batteries,” IEEE Transactions on Smart Grid, 9 (5), pp. 4726–4734, 2018.
- F. Farokhi, H. Sandberg,“Ensuring Privacy with Constrained Additive Noise by Minimizing Fisher Information,” Automatica, Vol 99, pp. 275–288, 2019.
- F. Farokhi, M. Egerstedt “Optimal Stochastic Evasive Maneuvers Using the Schrödinger’s Equation,” IEEE Control Systems Letters, 3 (3), pp. 517–522, 2019.
- F. Farokhi, “A Fundamental Bound on Performance of Non-Intrusive Load Monitoring with Application to Smart Meter Privacy,” Submitted.
- F. Farokhi, H. Sandberg, “Fisher Information Privacy with Application to Smart Meter Privacy Using HVAC Units,” in F. Farokhi, (Eds.), Privacy of Dynamical Systems, Internet of Things, Springer, 2020.
- F. Farokhi, “Taking a Lesson from Quantum Particles for Statistical Data Privacy,” arXiv:1908.04954 [cs.CR], 2019.
Collaborators: Henrik Sandberg, Magnus Egerstedt
Funding: veski Victoria Fellowship, The University of Melbourne