English

A bounded-noise mechanism for differential privacy

Data Structures and Algorithms 2021-11-09 v2 Cryptography and Security Machine Learning

Abstract

We present an asymptotically optimal (ϵ,δ)(\epsilon,\delta) differentially private mechanism for answering multiple, adaptively asked, Δ\Delta-sensitive queries, settling the conjecture of Steinke and Ullman [2020]. Our algorithm has a significant advantage that it adds independent bounded noise to each query, thus providing an absolute error bound. Additionally, we apply our algorithm in adaptive data analysis, obtaining an improved guarantee for answering multiple queries regarding some underlying distribution using a finite sample. Numerical computations show that the bounded-noise mechanism outperforms the Gaussian mechanism in many standard settings.

Keywords

Cite

@article{arxiv.2012.03817,
  title  = {A bounded-noise mechanism for differential privacy},
  author = {Yuval Dagan and Gil Kur},
  journal= {arXiv preprint arXiv:2012.03817},
  year   = {2021}
}
R2 v1 2026-06-23T20:47:13.323Z