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. 2302.05552
20
8

Algorithmically Effective Differentially Private Synthetic Data

11 February 2023
Yi He
Roman Vershynin
Yizhe Zhu
    SyDa
ArXivPDFHTML
Abstract

We present a highly effective algorithmic approach for generating ε\varepsilonε-differentially private synthetic data in a bounded metric space with near-optimal utility guarantees under the 1-Wasserstein distance. In particular, for a dataset XXX in the hypercube [0,1]d[0,1]^d[0,1]d, our algorithm generates synthetic dataset YYY such that the expected 1-Wasserstein distance between the empirical measure of XXX and YYY is O((εn)−1/d)O((\varepsilon n)^{-1/d})O((εn)−1/d) for d≥2d\geq 2d≥2, and is O(log⁡2(εn)(εn)−1)O(\log^2(\varepsilon n)(\varepsilon n)^{-1})O(log2(εn)(εn)−1) for d=1d=1d=1. The accuracy guarantee is optimal up to a constant factor for d≥2d\geq 2d≥2, and up to a logarithmic factor for d=1d=1d=1. Our algorithm has a fast running time of O(εdn)O(\varepsilon dn)O(εdn) for all d≥1d\geq 1d≥1 and demonstrates improved accuracy compared to the method in (Boedihardjo et al., 2022) for d≥2d\geq 2d≥2.

View on arXiv
Comments on this paper