Test without Trust: Optimal Locally Private Distribution Testing

Abstract

We study the problem of distribution testing when the samples can only be accessed using a locally differentially private mechanism and focus on two representative testing questions of identity (goodness-of-fit) and independence testing for discrete distributions. We are concerned with two settings: First, when we insist on using an already deployed, general-purpose locally differentially private mechanism such as the popular RAPPOR or the recently introduced Hadamard Response for collecting data, and must build our tests based on the data collected via this mechanism; and second, when no such restriction is imposed, and we can design a bespoke mechanism specifically for testing. For the latter purpose, we introduce the Randomized Aggregated Private Testing Optimal Response (RAPTOR) mechanism which is remarkably simple and requires only one bit of communication per sample. We propose tests based on these mechanisms and analyze their sample complexities. Each proposed test can be implemented efficiently. In each case (barring one), we complement our performance bounds for algorithms with information-theoretic lower bounds and establish sample optimality of our proposed algorithm. A peculiar feature that emerges is that our sample-optimal algorithm based on RAPTOR uses public-coins, and any test based on RAPPOR or Hadamard Response, which are both private-coin mechanisms, requires significantly more samples.

Cited by in corpus (17)

• A Comprehensive Survey on Local Differential Privacy Toward Data Statistics and Analysis
• On the Power of Multiple Anonymous Messages
• Locally Private Hypothesis Selection
• Locally private non-asymptotic testing of discrete distributions is faster using interactive mechanisms
• Distributed Simulation and Distributed Inference
• Private Identity Testing for High-Dimensional Distributions
• Differentially Private Assouad, Fano, and Le Cam
• Connecting Robust Shuffle Privacy and Pan-Privacy
• Context-Aware Local Differential Privacy
• Pan-Private Uniformity Testing
• Robust Testing and Estimation under Manipulation Attacks
• Domain Compression and its Application to Randomness-Optimal Distributed Goodness-of-Fit
• Statistical Inference in the Differential Privacy Model
• Differential Privacy in the Shuffle Model: A Survey of Separations
• Inference under Information Constraints III: Local Privacy Constraints
• Communication Complexity of Distributed High Dimensional Correlation Testing
• Exponential Separations in Local Differential Privacy