ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2203.01537
14
11

Private High-Dimensional Hypothesis Testing

3 March 2022
Shyam Narayanan
    FedML
ArXivPDFHTML
Abstract

We provide improved differentially private algorithms for identity testing of high-dimensional distributions. Specifically, for ddd-dimensional Gaussian distributions with known covariance Σ\SigmaΣ, we can test whether the distribution comes from N(μ∗,Σ)\mathcal{N}(\mu^*, \Sigma)N(μ∗,Σ) for some fixed μ∗\mu^*μ∗ or from some N(μ,Σ)\mathcal{N}(\mu, \Sigma)N(μ,Σ) with total variation distance at least α\alphaα from N(μ∗,Σ)\mathcal{N}(\mu^*, \Sigma)N(μ∗,Σ) with (ε,0)(\varepsilon, 0)(ε,0)-differential privacy, using only \[\tilde{O}\left(\frac{d^{1/2}}{\alpha^2} + \frac{d^{1/3}}{\alpha^{4/3} \cdot \varepsilon^{2/3}} + \frac{1}{\alpha \cdot \varepsilon}\right)\] samples if the algorithm is allowed to be computationally inefficient, and only \[\tilde{O}\left(\frac{d^{1/2}}{\alpha^2} + \frac{d^{1/4}}{\alpha \cdot \varepsilon}\right)\] samples for a computationally efficient algorithm. We also provide a matching lower bound showing that our computationally inefficient algorithm has optimal sample complexity. We also extend our algorithms to various related problems, including mean testing of Gaussians with bounded but unknown covariance, uniformity testing of product distributions over {−1,1}d\{-1, 1\}^d{−1,1}d, and tolerant testing. Our results improve over the previous best work of Canonne et al.~\cite{CanonneKMUZ20} for both computationally efficient and inefficient algorithms, and even our computationally efficient algorithm matches the optimal \emph{non-private} sample complexity of O(dα2)O\left(\frac{\sqrt{d}}{\alpha^2}\right)O(α2d​​) in many standard parameter settings. In addition, our results show that, surprisingly, private identity testing of ddd-dimensional Gaussians can be done with fewer samples than private identity testing of discrete distributions over a domain of size ddd \cite{AcharyaSZ18}, which refutes a conjectured lower bound of~\cite{CanonneKMUZ20}.

View on arXiv
Comments on this paper