Private Confidence Sets

Abstract

We consider statistical inference under privacy constraints. In particular, we give differentially private algorithms for estimating coverage probabilities and computing valid confidence sets, and prove upper bounds on the error of our estimates and the length of our confidence sets. Our bounds apply to broad classes of data distributions and statistics of interest, and for fixed $\varepsilon$ we match the higher-order asymptotic accuracy of the standard (non-private) non-parametric bootstrap.

PDF NeurIPSW OpenReview Semantic Scholar

Cite

Text

Chadha et al. "Private Confidence Sets." NeurIPS 2021 Workshops: PRIML, 2021.

Markdown

[Chadha et al. "Private Confidence Sets." NeurIPS 2021 Workshops: PRIML, 2021.](https://mlanthology.org/neuripsw/2021/chadha2021neuripsw-private/)

BibTeX

@inproceedings{chadha2021neuripsw-private,
  title     = {{Private Confidence Sets}},
  author    = {Chadha, Karan and Duchi, John and Kuditipudi, Rohith},
  booktitle = {NeurIPS 2021 Workshops: PRIML},
  year      = {2021},
  url       = {https://mlanthology.org/neuripsw/2021/chadha2021neuripsw-private/}
}