Lightweight Protocols for Distributed Private Quantile Estimation

Anders Aamand, Fabrizio Boninsegna, Abigail Gentle, Jacob Imola, Rasmus Pagh
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:27-58, 2025.

Abstract

Distributed data analysis is a large and growing field driven by a massive proliferation of user devices, and by privacy concerns surrounding the centralised storage of data. We consider two adaptive algorithms for estimating one quantile (e.g. the median) when each user holds a single data point lying in a domain $[B]$ that can be queried once through a private mechanism; one under local differential privacy (LDP) and another for shuffle differential privacy (shuffle-DP). In the adaptive setting we present an $\varepsilon$-LDP algorithm which can estimate any quantile within error $\alpha$ only requiring $O(\frac{\log B}{\varepsilon^2\alpha^2})$ users, and an $(\varepsilon,\delta)$-shuffle DP algorithm requiring only $\widetilde{O}((\frac{1}{\varepsilon^2}+\frac{1}{\alpha^2})\log B)$ users. Prior (nonadaptive) algorithms require more users by several logarithmic factors in $B$. We further provide a matching lower bound for adaptive protocols, showing that our LDP algorithm is optimal in the low-$\varepsilon$ regime. Additionally, we establish lower bounds against non-adaptive protocols which paired with our understanding of the adaptive case, proves a fundamental separation between these models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-aamand25a, title = {Lightweight Protocols for Distributed Private Quantile Estimation}, author = {Aamand, Anders and Boninsegna, Fabrizio and Gentle, Abigail and Imola, Jacob and Pagh, Rasmus}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {27--58}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/aamand25a/aamand25a.pdf}, url = {https://proceedings.mlr.press/v267/aamand25a.html}, abstract = {Distributed data analysis is a large and growing field driven by a massive proliferation of user devices, and by privacy concerns surrounding the centralised storage of data. We consider two adaptive algorithms for estimating one quantile (e.g. the median) when each user holds a single data point lying in a domain $[B]$ that can be queried once through a private mechanism; one under local differential privacy (LDP) and another for shuffle differential privacy (shuffle-DP). In the adaptive setting we present an $\varepsilon$-LDP algorithm which can estimate any quantile within error $\alpha$ only requiring $O(\frac{\log B}{\varepsilon^2\alpha^2})$ users, and an $(\varepsilon,\delta)$-shuffle DP algorithm requiring only $\widetilde{O}((\frac{1}{\varepsilon^2}+\frac{1}{\alpha^2})\log B)$ users. Prior (nonadaptive) algorithms require more users by several logarithmic factors in $B$. We further provide a matching lower bound for adaptive protocols, showing that our LDP algorithm is optimal in the low-$\varepsilon$ regime. Additionally, we establish lower bounds against non-adaptive protocols which paired with our understanding of the adaptive case, proves a fundamental separation between these models.} }
Endnote
%0 Conference Paper %T Lightweight Protocols for Distributed Private Quantile Estimation %A Anders Aamand %A Fabrizio Boninsegna %A Abigail Gentle %A Jacob Imola %A Rasmus Pagh %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-aamand25a %I PMLR %P 27--58 %U https://proceedings.mlr.press/v267/aamand25a.html %V 267 %X Distributed data analysis is a large and growing field driven by a massive proliferation of user devices, and by privacy concerns surrounding the centralised storage of data. We consider two adaptive algorithms for estimating one quantile (e.g. the median) when each user holds a single data point lying in a domain $[B]$ that can be queried once through a private mechanism; one under local differential privacy (LDP) and another for shuffle differential privacy (shuffle-DP). In the adaptive setting we present an $\varepsilon$-LDP algorithm which can estimate any quantile within error $\alpha$ only requiring $O(\frac{\log B}{\varepsilon^2\alpha^2})$ users, and an $(\varepsilon,\delta)$-shuffle DP algorithm requiring only $\widetilde{O}((\frac{1}{\varepsilon^2}+\frac{1}{\alpha^2})\log B)$ users. Prior (nonadaptive) algorithms require more users by several logarithmic factors in $B$. We further provide a matching lower bound for adaptive protocols, showing that our LDP algorithm is optimal in the low-$\varepsilon$ regime. Additionally, we establish lower bounds against non-adaptive protocols which paired with our understanding of the adaptive case, proves a fundamental separation between these models.
APA
Aamand, A., Boninsegna, F., Gentle, A., Imola, J. & Pagh, R.. (2025). Lightweight Protocols for Distributed Private Quantile Estimation. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:27-58 Available from https://proceedings.mlr.press/v267/aamand25a.html.

Related Material