Graphical-model based estimation and inference for differential privacy

Ryan Mckenna, Daniel Sheldon, Gerome Miklau
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:4435-4444, 2019.

Abstract

Many privacy mechanisms reveal high-level information about a data distribution through noisy measurements. It is common to use this information to estimate the answers to new queries. In this work, we provide an approach to solve this estimation problem efficiently using graphical models, which is particularly effective when the distribution is high-dimensional but the measurements are over low-dimensional marginals. We show that our approach is far more efficient than existing estimation techniques from the privacy literature and that it can improve the accuracy and scalability of many state-of-the-art mechanisms.

Cite this Paper


BibTeX
@InProceedings{pmlr-v97-mckenna19a, title = {Graphical-model based estimation and inference for differential privacy}, author = {Mckenna, Ryan and Sheldon, Daniel and Miklau, Gerome}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {4435--4444}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/mckenna19a/mckenna19a.pdf}, url = {https://proceedings.mlr.press/v97/mckenna19a.html}, abstract = {Many privacy mechanisms reveal high-level information about a data distribution through noisy measurements. It is common to use this information to estimate the answers to new queries. In this work, we provide an approach to solve this estimation problem efficiently using graphical models, which is particularly effective when the distribution is high-dimensional but the measurements are over low-dimensional marginals. We show that our approach is far more efficient than existing estimation techniques from the privacy literature and that it can improve the accuracy and scalability of many state-of-the-art mechanisms.} }
Endnote
%0 Conference Paper %T Graphical-model based estimation and inference for differential privacy %A Ryan Mckenna %A Daniel Sheldon %A Gerome Miklau %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-mckenna19a %I PMLR %P 4435--4444 %U https://proceedings.mlr.press/v97/mckenna19a.html %V 97 %X Many privacy mechanisms reveal high-level information about a data distribution through noisy measurements. It is common to use this information to estimate the answers to new queries. In this work, we provide an approach to solve this estimation problem efficiently using graphical models, which is particularly effective when the distribution is high-dimensional but the measurements are over low-dimensional marginals. We show that our approach is far more efficient than existing estimation techniques from the privacy literature and that it can improve the accuracy and scalability of many state-of-the-art mechanisms.
APA
Mckenna, R., Sheldon, D. & Miklau, G.. (2019). Graphical-model based estimation and inference for differential privacy. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:4435-4444 Available from https://proceedings.mlr.press/v97/mckenna19a.html.

Related Material