Privacy-Preserving Constrained Quadratic Optimization With Fisher Information

• Published in: IEEE signal processing letters Vol. 27; pp. 545 - 549
• Main Author: Farokhi, Farhad
• Format: Journal Article
• Language: English
• Published: New York IEEE 2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; constrained optimization; Constraints; Cost function; Differential privacy; Fisher information; Noise; Noise measurement; noisy gradient descent; Optimization; Optimization algorithm; Parameter estimation; Performance degradation; Privacy; privacy-preserving optimization; quadratic cost; Signal processing algorithms

Noisy (stochastic) gradient descent is used to develop privacy-preserving algorithms for solving constrained quadratic optimization problems. The variance of the error of an adversary's estimate of the parameters of the quadratic cost function based on iterates of the algorithm is related to th...