Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Quantifying the privacy risks of learning high-dimensional graphical models

Murakonda, Sasi Kumar, Shokri, Reza and Theodorakopoulos, Georgios ORCID: 2021. Quantifying the privacy risks of learning high-dimensional graphical models. Presented at: 24th International Conference on Artificial Intelligence and Statistics (AISTATS), virtual, 13-15 April 2021. Proceedings of Machine Learning Research. pp. 2287-2295.

[thumbnail of kumar-murakonda21a.pdf]
PDF - Published Version
Download (491kB) | Preview


Models leak information about their training data. This enables attackers to infer sensitive information about their training sets, notably determine if a data sample was part of the model’s training set. The existing works empirically show the possibility of these membership inference (tracing) attacks against complex deep learning models. However, the attack results are dependent on the specific training data, can be obtained only after the tedious process of training the model and performing the attack, and are missing any measure of the confidence and unused potential power of the attack. In this paper, we theoretically analyze the maximum power of tracing attacks against high-dimensional graphical models, with the focus on Bayesian networks. We provide a tight upper bound on the power (true positive rate) of these attacks, with respect to their error (false positive rate), for a given model structure even before learning its parameters. As it should be, the bound is independent of the knowledge and algorithm of any specific attack. It can help in identifying which model structures leak more information, how adding new parameters to the model increases its privacy risk, and what can be gained by adding new data points to decrease the overall information leakage. It provides a measure of the potential leakage of a model given its structure, as a function of the model complexity and the size of the training set.

Item Type: Conference or Workshop Item (Paper)
Status: Published
Schools: Computer Science & Informatics
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
ISSN: 2640-3498
Date of First Compliant Deposit: 22 April 2021
Date of Acceptance: 23 January 2021
Last Modified: 09 Nov 2022 10:47

Actions (repository staff only)

Edit Item Edit Item


Downloads per month over past year

View more statistics