Dropout-Robust Mechanisms for Differentially Private and Fully Decentralized Mean Estimation
This addresses the challenge of privacy-preserving data processing in decentralized networks, offering a robust solution for applications like distributed analytics, though it is incremental in improving existing decentralized techniques.
The paper tackles the problem of achieving differentially private mean estimation in fully decentralized settings, proposing the IncA protocol which uses low-variance correlated noise to maintain accuracy comparable to centralized methods and reduce accuracy loss from dropouts.
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.