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

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2002.09465
134
29
v1v2 (latest)

Locally Private Hypothesis Selection

21 February 2020
Sivakanth Gopi
Gautam Kamath
Janardhan Kulkarni
Aleksandar Nikolov
Zhiwei Steven Wu
Huanyu Zhang
ArXiv (abs)PDFHTML
Abstract

We initiate the study of hypothesis selection under local differential privacy. Given samples from an unknown probability distribution ppp and a set of kkk probability distributions Q\mathcal{Q}Q, we aim to output, under the constraints of ε\varepsilonε-local differential privacy, a distribution from Q\mathcal{Q}Q whose total variation distance to ppp is comparable to the best such distribution. This is a generalization of the classic problem of kkk-wise simple hypothesis testing, which corresponds to when p∈Qp \in \mathcal{Q}p∈Q, and we wish to identify ppp. Absent privacy constraints, this problem requires O(log⁡k)O(\log k)O(logk) samples from ppp, and it was recently shown that the same complexity is achievable under (central) differential privacy. However, the naive approach to this problem under local differential privacy would require O~(k2)\tilde O(k^2)O~(k2) samples. We first show that the constraint of local differential privacy incurs an exponential increase in cost: any algorithm for this problem requires at least Ω(k)\Omega(k)Ω(k) samples. Second, for the special case of kkk-wise simple hypothesis testing, we provide a non-interactive algorithm which nearly matches this bound, requiring O~(k)\tilde O(k)O~(k) samples. Finally, we provide sequentially interactive algorithms for the general case, requiring O~(k)\tilde O(k)O~(k) samples and only O(log⁡log⁡k)O(\log \log k)O(loglogk) rounds of interactivity. Our algorithms are achieved through a reduction to maximum selection with adversarial comparators, a problem of independent interest for which we initiate study in the parallel setting. For this problem, we provide a family of algorithms for each number of allowed rounds of interaction ttt, as well as lower bounds showing that they are near-optimal for every ttt. Notably, our algorithms result in exponential improvements on the round complexity of previous methods.

View on arXiv
Comments on this paper