Statistical Approximating Distributions Under Differential Privacy

Statistics computed from data are viewed as random variables. When they are used for tasks like hypothesis testing and confidence intervals, their true finite sample distributions are often replaced by approximating distributions that are easier to work with (for example, the Gaussian, which results from using approximations justified by the Central Limit Theorem). When data are perturbed by differential privacy, the approximating distributions also need to be modified. Prior work provided various competing methods for creating such approximating distributions with little formal justification beyond the fact that they worked well empirically. In this paper, we study the question of how to generate statistical approximating distributions for differentially private statistics, provide finite sample guarantees for the quality of the approximations.

• Publication year

  2018

• Venue

  Journal of Privacy and Confidentiality

• Publication date

  2018-12-21

• Fields of study

  Mathematics, Computer Science

• Identifiers
  DOI 10.29012/JPC.666
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• A Probabilistic Procedure for Anonymisation and Analysis of Perturbed Datasets

  2018cited by this paper
• Confidentiality and Differential Privacy in the Dissemination of Frequency Tables

  2018cited by this paper
• Differentially Private Uniformly Most Powerful Tests for Binomial Data

  2018cited by this paper
• Priv'IT: Private and Sample Efficient Identity Testing

  2017cited by this paper
• Finite Sample Differentially Private Confidence Intervals

  2017cited by this paper
• Differentially Private Testing of Identity and Closeness of Discrete Distributions

  2017cited by this paper
• A New Class of Private Chi-Square Hypothesis Tests

  2017influential reference
• Differentially Private Chi-Squared Hypothesis Testing: Goodness of Fit and Independence Testing

  2016influential reference
• Differentially Private Regression Diagnostics

  2016cited by this paper
• Minimax Optimal Procedures for Locally Private Estimation

  2016cited by this paper
• Concentrated Differential Privacy: Simplifications, Extensions, and Lower Bounds

  2016influential reference
• Differentially Private Ordinary Least Squares: $t$-Values, Confidence Intervals and Rejecting Null-Hypotheses

  2015cited by this paper
• Private False Discovery Rate Control

  2015cited by this paper
• Designing statistical privacy for your data

  2015cited by this paper
• Differentially Private Hypothesis Testing, Revisited

  2015cited by this paper
• Stein’s method for functions of multivariate normal random variables

  2015influential reference
• Differential Privacy for Social Science Inference

  2015cited by this paper
• Scalable privacy-preserving data sharing methodology for genome-wide association studies

  2014influential reference
• Differentially Private Hypothesis Testing For Normal Random Variables.

  2014influential reference
• Privacy-preserving data exploration in genome-wide association studies

  2013influential reference
• Convergence Rates for Differentially Private Statistical Estimation

  2012cited by this paper
• Privacy-Preserving Data Sharing for Genome-Wide Association Studies

  2012influential reference
• Fundamentals of Stein's method

  2011cited by this paper
• Privacy-preserving statistical estimation with optimal convergence rates

  2011influential reference
• How Can We Analyze Differentially-Private Synthetic Datasets?

  2011influential reference
• On the Kantorovich–Rubinstein theorem

  2011influential reference
• Differential Privacy and the Risk-Utility Tradeoff for Multi-dimensional Contingency Tables

  2010cited by this paper
• Differential Privacy for Clinical Trial Data: Preliminary Evaluations

  2009cited by this paper
• Differential privacy and robust statistics

  2009cited by this paper
• A Statistical Framework for Differential Privacy

  2008cited by this paper
• Calibrating Noise to Sensitivity in Private Data Analysis

  2006influential reference
• A Course in Large Sample Theory

  1996influential reference
• A bound for the error in the normal approximation to the distribution of a sum of dependent random variables

  1972cited by this paper
• The accuracy of the Gaussian approximation to the sum of independent variates

  1941cited by this paper

• Near-Optimal Private Tests for Simple and MLR Hypotheses

  2026cites this paper
• Black-Box Differentially Private Nonparametric Confidence Intervals Under Minimal Assumptions

  2025cites this paper
• Fiducial Matching: Differentially Private Inference for Categorical Data

  2025cites this paper
• Particle Filter for Bayesian Inference on Privatized Data

  2025cites this paper
• Statistical Inference for Privatized Data with Unknown Sample Size

  2024cites this paper
• Simulation-Based, Finite-Sample Inference for Privatized Data

  2023influential citation
• Differentially Private Conditional Independence Testing

  2023cites this paper
• The Test of Tests: A Framework For Differentially Private Hypothesis Testing

  2023cites this paper
• Statistical Data Privacy: A Song of Privacy and Utility

  2022cites this paper
• Data Augmentation MCMC for Bayesian Inference from Privatized Data

  2022cites this paper
• Differentially Private Kolmogorov-Smirnov-Type Tests

  2022influential citation
• Adjusting Queries to Statistical Procedures Under Differential Privacy

  2021cites this paper
• Canonical Noise Distributions and Private Hypothesis Tests

  2021cites this paper
• A Primer on Private Statistics

  2020cites this paper
• Parametric Bootstrap for Differentially Private Confidence Intervals

  2020cites this paper
• One Step to Efficient Synthetic Data

  2020cites this paper
• Differentially Private Inference for Binomial Data

  2019influential citation
• Benefits and Pitfalls of the Exponential Mechanism with Applications to Hilbert Spaces and Functional PCA

  2019cites this paper
• Private Identity Testing for High-Dimensional Distributions

  2019cites this paper
• Elliptical Perturbations for Differential Privacy

  2019cites this paper
• The structure of optimal private tests for simple hypotheses

  2018cites this paper