Measuring data leakage in machine-learning models with Fisher information

Awni Hannun, Chuan Guo, Laurens van der Maaten
Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, PMLR 161:760-770, 2021.

Abstract

Machine-learning models contain information about the data they were trained on. This information leaks either through the model itself or through predictions made by the model. Consequently, when the training data contains sensitive attributes, assessing the amount of information leakage is paramount. We propose a method to quantify this leakage using the Fisher information of the model about the data. Unlike the worst-case a priori guarantees of differential privacy, Fisher information loss measures leakage with respect to specific examples, attributes, or sub-populations within the dataset. We motivate Fisher information loss through the Cram\’{e}r-Rao bound and delineate the implied threat model. We provide efficient methods to compute Fisher information loss for output-perturbed generalized linear models. Finally, we empirically validate Fisher information loss as a useful measure of information leakage.

Cite this Paper

BibTeX
@InProceedings{pmlr-v161-hannun21a, title = {Measuring data leakage in machine-learning models with Fisher information}, author = {Hannun, Awni and Guo, Chuan and van der Maaten, Laurens}, booktitle = {Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence}, pages = {760--770}, year = {2021}, editor = {de Campos, Cassio and Maathuis, Marloes H.}, volume = {161}, series = {Proceedings of Machine Learning Research}, month = {27--30 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v161/hannun21a/hannun21a.pdf}, url = {https://proceedings.mlr.press/v161/hannun21a.html}, abstract = {Machine-learning models contain information about the data they were trained on. This information leaks either through the model itself or through predictions made by the model. Consequently, when the training data contains sensitive attributes, assessing the amount of information leakage is paramount. We propose a method to quantify this leakage using the Fisher information of the model about the data. Unlike the worst-case a priori guarantees of differential privacy, Fisher information loss measures leakage with respect to specific examples, attributes, or sub-populations within the dataset. We motivate Fisher information loss through the Cram\’{e}r-Rao bound and delineate the implied threat model. We provide efficient methods to compute Fisher information loss for output-perturbed generalized linear models. Finally, we empirically validate Fisher information loss as a useful measure of information leakage.} }
Endnote
%0 Conference Paper %T Measuring data leakage in machine-learning models with Fisher information %A Awni Hannun %A Chuan Guo %A Laurens van der Maaten %B Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2021 %E Cassio de Campos %E Marloes H. Maathuis %F pmlr-v161-hannun21a %I PMLR %P 760--770 %U https://proceedings.mlr.press/v161/hannun21a.html %V 161 %X Machine-learning models contain information about the data they were trained on. This information leaks either through the model itself or through predictions made by the model. Consequently, when the training data contains sensitive attributes, assessing the amount of information leakage is paramount. We propose a method to quantify this leakage using the Fisher information of the model about the data. Unlike the worst-case a priori guarantees of differential privacy, Fisher information loss measures leakage with respect to specific examples, attributes, or sub-populations within the dataset. We motivate Fisher information loss through the Cram\’{e}r-Rao bound and delineate the implied threat model. We provide efficient methods to compute Fisher information loss for output-perturbed generalized linear models. Finally, we empirically validate Fisher information loss as a useful measure of information leakage.
APA
Hannun, A., Guo, C. & van der Maaten, L.. (2021). Measuring data leakage in machine-learning models with Fisher information. Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 161:760-770 Available from https://proceedings.mlr.press/v161/hannun21a.html.

Related Material