Privacy-Preserving Distributed Processing: Metrics, Bounds and Algorithms

Privacy-preserving distributed processing has recently attracted considerable attention. It aims to design solutions for conducting signal tasks over networks in a decentralized fashion without violating privacy. Many existing algorithms can be adopted solve this problem such as differential privacy, secure multiparty computation, and the proposed optimization based subspace perturbation algorithms. However, since each of them is derived from different context metrics assumptions, it hard choose or an appropriate algorithm processing. In order address problem, we first propose general mutual information information-theoretical that are able compare relate these terms two key aspects: output utility individual We consider widely-used adversary models, passive eavesdropping adversary. Moreover, derive lower bound on privacy which helps understand nature provides insights preferred given conditions. To validate above claims, investigate concrete example number state-of-the-art approaches concerned aspects using not only theoretical analysis but also numerical validation. Finally, discuss provide principles designing applications.