Pain-Free Random Differential Privacy with Sensitivity Sampling

Benjamin I. P. Rubinstein, Francesco Aldà
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:2950-2959, 2017.

Abstract

Popular approaches to differential privacy, such as the Laplace and exponential mechanisms, calibrate randomised smoothing through global sensitivity of the target non-private function. Bounding such sensitivity is often a prohibitively complex analytic calculation. As an alternative, we propose a straightforward sampler for estimating sensitivity of non-private mechanisms. Since our sensitivity estimates hold with high probability, any mechanism that would be $(\epsilon,\delta)$-differentially private under bounded global sensitivity automatically achieves $(\epsilon,\delta,\gamma)$-random differential privacy (Hall et al. 2012), without any target-specific calculations required. We demonstrate on worked example learners how our usable approach adopts a naturally-relaxed privacy guarantee, while achieving more accurate releases even for non-private functions that are black-box computer programs.

Cite this Paper


BibTeX
@InProceedings{pmlr-v70-rubinstein17a, title = {Pain-Free Random Differential Privacy with Sensitivity Sampling}, author = {Benjamin I. P. Rubinstein and Francesco Ald{\`a}}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {2950--2959}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/rubinstein17a/rubinstein17a.pdf}, url = {https://proceedings.mlr.press/v70/rubinstein17a.html}, abstract = {Popular approaches to differential privacy, such as the Laplace and exponential mechanisms, calibrate randomised smoothing through global sensitivity of the target non-private function. Bounding such sensitivity is often a prohibitively complex analytic calculation. As an alternative, we propose a straightforward sampler for estimating sensitivity of non-private mechanisms. Since our sensitivity estimates hold with high probability, any mechanism that would be $(\epsilon,\delta)$-differentially private under bounded global sensitivity automatically achieves $(\epsilon,\delta,\gamma)$-random differential privacy (Hall et al. 2012), without any target-specific calculations required. We demonstrate on worked example learners how our usable approach adopts a naturally-relaxed privacy guarantee, while achieving more accurate releases even for non-private functions that are black-box computer programs.} }
Endnote
%0 Conference Paper %T Pain-Free Random Differential Privacy with Sensitivity Sampling %A Benjamin I. P. Rubinstein %A Francesco Aldà %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-rubinstein17a %I PMLR %P 2950--2959 %U https://proceedings.mlr.press/v70/rubinstein17a.html %V 70 %X Popular approaches to differential privacy, such as the Laplace and exponential mechanisms, calibrate randomised smoothing through global sensitivity of the target non-private function. Bounding such sensitivity is often a prohibitively complex analytic calculation. As an alternative, we propose a straightforward sampler for estimating sensitivity of non-private mechanisms. Since our sensitivity estimates hold with high probability, any mechanism that would be $(\epsilon,\delta)$-differentially private under bounded global sensitivity automatically achieves $(\epsilon,\delta,\gamma)$-random differential privacy (Hall et al. 2012), without any target-specific calculations required. We demonstrate on worked example learners how our usable approach adopts a naturally-relaxed privacy guarantee, while achieving more accurate releases even for non-private functions that are black-box computer programs.
APA
Rubinstein, B.I.P. & Aldà, F.. (2017). Pain-Free Random Differential Privacy with Sensitivity Sampling. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:2950-2959 Available from https://proceedings.mlr.press/v70/rubinstein17a.html.

Related Material