Differentially Private Online Community Detection for Censored Block Models: Algorithms and Fundamental Limits

TL;DR

This paper develops algorithms for private online community detection in dynamic networks modeled by censored block models, analyzing the tradeoffs between privacy, detection delay, and community recovery with theoretical guarantees.

Contribution

It introduces joint change detection and community estimation methods under edge differential privacy, providing necessary and sufficient conditions for their effectiveness.

Findings

Proposed algorithms achieve privacy-preserving community detection.

Theoretical bounds characterize tradeoffs between privacy and detection accuracy.

Validated methods with simulations and real data examples.

Abstract

We study the private online change detection problem for dynamic communities, using a censored block model (CBM). We consider edge differential privacy (DP) in both local and central settings, and propose joint change detection and community estimation procedures for both scenarios. We seek to understand the fundamental tradeoffs between the privacy budget, detection delay, and exact community recovery of community labels. Further, we provide theoretical guarantees for the effectiveness of our proposed method by showing necessary and sufficient conditions for change detection and exact recovery under edge DP. Simulation and real data examples are provided to validate the proposed methods.