Recursive Binary Identification with Differential Privacy and Data Tampering Attacks
Jimin Wang, Jieming Ke, Jin Guo, Yanlong Zhao
TL;DR
The paper addresses parameter identification in bandwidth-limited and insecure wireless settings where measurements are quantized to a single bit. It couples differential privacy with data tampering attacks and presents a recursive projection algorithm that achieves almost sure and mean-square convergence under these constraints. A detailed convergence and privacy analysis shows that the mean-square error decays as $O(1/k)$ both with and without privacy, indicating negligible asymptotic cost for privacy guarantees. The framework extends to distributed multi-agent networks, preserving convergence guarantees under connectivity, with simulations illustrating robustness to tampering and privacy noise in both single-agent and multi-agent scenarios.
Abstract
In this paper, we consider the parameter estimation in a bandwidth-constrained sensor network communicating through an insecure medium. The sensor performs a local quantization, and transmits a 1-bit message to an estimation center through a wireless medium where the transmission of information is vulnerable to attackers. Both eavesdroppers and data tampering attackers are considered in our setting. A differential privacy method is used to protect the sensitive information against eavesdroppers. Then, a recursive projection algorithm is proposed such that the estimation center achieves the almost sure convergence and mean-square convergence when quantized measurements, differential privacy, and data tampering attacks are considered in a uniform framework. A privacy analysis including the convergence rate with privacy or without privacy is given. Further, we extend the problem to multi-agent systems. For this case, a distributed recursive projection algorithm is proposed with guaranteed almost sure and mean square convergence. A simulation example is provided to illustrate the effectiveness of the proposed algorithms.