Optimal Differentially Private Sampling of Unbounded Gaussians

Valentio Iverson, Gautam Kamath, Argyris Mouzakis
Proceedings of Thirty Eighth Conference on Learning Theory, PMLR 291:2893-2941, 2025.

Abstract

We provide the first $\widetilde{\mathcal{O}}(d)$-sample algorithm for sampling from unbounded Gaussian distributions under the constraint of $(\varepsilon, \delta)$-differential privacy. This is a quadratic improvement over previous results for the same problem, settling an open question of Ghazi, Hu, Kumar, and Manurangsi.

Cite this Paper

BibTeX
@InProceedings{pmlr-v291-iverson25a, title = {Optimal Differentially Private Sampling of Unbounded Gaussians}, author = {Iverson, Valentio and Kamath, Gautam and Mouzakis, Argyris}, booktitle = {Proceedings of Thirty Eighth Conference on Learning Theory}, pages = {2893--2941}, year = {2025}, editor = {Haghtalab, Nika and Moitra, Ankur}, volume = {291}, series = {Proceedings of Machine Learning Research}, month = {30 Jun--04 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v291/main/assets/iverson25a/iverson25a.pdf}, url = {https://proceedings.mlr.press/v291/iverson25a.html}, abstract = {We provide the first $\widetilde{\mathcal{O}}(d)$-sample algorithm for sampling from unbounded Gaussian distributions under the constraint of $(\varepsilon, \delta)$-differential privacy. This is a quadratic improvement over previous results for the same problem, settling an open question of Ghazi, Hu, Kumar, and Manurangsi.} }
Endnote
%0 Conference Paper %T Optimal Differentially Private Sampling of Unbounded Gaussians %A Valentio Iverson %A Gautam Kamath %A Argyris Mouzakis %B Proceedings of Thirty Eighth Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2025 %E Nika Haghtalab %E Ankur Moitra %F pmlr-v291-iverson25a %I PMLR %P 2893--2941 %U https://proceedings.mlr.press/v291/iverson25a.html %V 291 %X We provide the first $\widetilde{\mathcal{O}}(d)$-sample algorithm for sampling from unbounded Gaussian distributions under the constraint of $(\varepsilon, \delta)$-differential privacy. This is a quadratic improvement over previous results for the same problem, settling an open question of Ghazi, Hu, Kumar, and Manurangsi.
APA
Iverson, V., Kamath, G. & Mouzakis, A.. (2025). Optimal Differentially Private Sampling of Unbounded Gaussians. Proceedings of Thirty Eighth Conference on Learning Theory, in Proceedings of Machine Learning Research 291:2893-2941 Available from https://proceedings.mlr.press/v291/iverson25a.html.

Related Material