Truncated Laplace and Gaussian mechanisms of RDP

TL;DR

This paper introduces truncated Laplace and Gaussian mechanisms that preserve Renyi Differential Privacy (RDP) while restricting outputs to specific intervals, improving utility and semantic validity in differential privacy applications.

Contribution

It proposes novel truncated mechanisms that maintain the same RDP as traditional ones, with explicit RDP expressions, enhancing privacy utility in bounded output scenarios.

Findings

Truncated mechanisms preserve RDP identical to untruncated ones.

Explicit RDP formulas for truncated Laplace and Gaussian mechanisms.

Potential for improved utility in differential privacy applications.

Abstract

The Laplace mechanism and the Gaussian mechanism are primary mechanisms in differential privacy, widely applicable to many scenarios involving numerical data. However, due to the infinite-range random variables they generate, the Laplace and Gaussian mechanisms may return values that are semantically impossible, such as negative numbers. To address this issue, we have designed the truncated Laplace mechanism and Gaussian mechanism. For a given truncation interval [a, b], the truncated Gaussian mechanism ensures the same Renyi Differential Privacy (RDP) as the untruncated mechanism, regardless of the values chosen for the truncation interval [a, b]. Similarly, the truncated Laplace mechanism, for specified interval [a, b], maintains the same RDP as the untruncated mechanism. We provide the RDP expressions for each of them. We believe that our study can further enhance the utility of differential privacy in specific applications.