Saved in:
Bibliographic Details
Main Authors: Thakkar, Janvi, Zizzo, Giulio, Maffeis, Sergio
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2401.10405
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911760787701760
author Thakkar, Janvi
Zizzo, Giulio
Maffeis, Sergio
author_facet Thakkar, Janvi
Zizzo, Giulio
Maffeis, Sergio
contents Malicious adversaries can attack machine learning models to infer sensitive information or damage the system by launching a series of evasion attacks. Although various work addresses privacy and security concerns, they focus on individual defenses, but in practice, models may undergo simultaneous attacks. This study explores the combination of adversarial training and differentially private training to defend against simultaneous attacks. While differentially-private adversarial training, as presented in DP-Adv, outperforms the other state-of-the-art methods in performance, it lacks formal privacy guarantees and empirical validation. Thus, in this work, we benchmark the performance of this technique using a membership inference attack and empirically show that the resulting approach is as private as non-robust private models. This work also highlights the need to explore privacy guarantees in dynamic training paradigms.
format Preprint
id arxiv_https___arxiv_org_abs_2401_10405
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Differentially Private and Adversarially Robust Machine Learning: An Empirical Evaluation
Thakkar, Janvi
Zizzo, Giulio
Maffeis, Sergio
Machine Learning
Malicious adversaries can attack machine learning models to infer sensitive information or damage the system by launching a series of evasion attacks. Although various work addresses privacy and security concerns, they focus on individual defenses, but in practice, models may undergo simultaneous attacks. This study explores the combination of adversarial training and differentially private training to defend against simultaneous attacks. While differentially-private adversarial training, as presented in DP-Adv, outperforms the other state-of-the-art methods in performance, it lacks formal privacy guarantees and empirical validation. Thus, in this work, we benchmark the performance of this technique using a membership inference attack and empirically show that the resulting approach is as private as non-robust private models. This work also highlights the need to explore privacy guarantees in dynamic training paradigms.
title Differentially Private and Adversarially Robust Machine Learning: An Empirical Evaluation
topic Machine Learning
url https://arxiv.org/abs/2401.10405