Dropout-Robust Mechanisms for Differentially Private and Fully Decentralized Mean Estimation

TL;DR

This paper introduces IncA, a fully decentralized, differentially private mean estimation protocol that maintains high accuracy even with network failures by using low-variance correlated noise and incremental sensitive information injection.

Contribution

The paper presents IncA, a novel decentralized protocol that achieves differential privacy with low communication overhead and robustness to dropouts, improving accuracy over existing methods.

Findings

IncA achieves centralized-level accuracy when no parties disconnect.

Correlated noise reduces accuracy loss during dropouts.

Empirical results confirm robustness and privacy guarantees.

Abstract

Achieving differentially private computations in decentralized settings poses significant challenges, particularly regarding accuracy, communication cost, and robustness against information leakage. While cryptographic solutions offer promise, they often suffer from high communication overhead or require centralization in the presence of network failures. Conversely, existing fully decentralized approaches typically rely on relaxed adversarial models or pairwise noise cancellation, the latter suffering from substantial accuracy degradation if parties unexpectedly disconnect. In this work, we propose IncA, a new protocol for fully decentralized mean estimation, a widely used primitive in data-intensive processing. Our protocol, which enforces differential privacy, requires no central orchestration and employs low-variance correlated noise, achieved by incrementally injecting sensitive information into the computation. First, we theoretically demonstrate that, when no parties permanently disconnect, our protocol achieves accuracy comparable to that of a centralized setting-already an improvement over most existing decentralized differentially private techniques. Second, we empirically show that our use of low-variance correlated noise significantly mitigates the accuracy loss experienced by existing techniques in the presence of dropouts.