The Relative Gaussian Mechanism and its Application to Private Gradient Descent

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Hendrikx, Hadrien, Mangold, Paul, Bellet, Aurélien
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914719227445248
author Hendrikx, Hadrien
Mangold, Paul
Bellet, Aurélien
author_facet Hendrikx, Hadrien
Mangold, Paul
Bellet, Aurélien
contents The Gaussian Mechanism (GM), which consists in adding Gaussian noise to a vector-valued query before releasing it, is a standard privacy protection mechanism. In particular, given that the query respects some L2 sensitivity property (the L2 distance between outputs on any two neighboring inputs is bounded), GM guarantees Rényi Differential Privacy (RDP). Unfortunately, precisely bounding the L2 sensitivity can be hard, thus leading to loose privacy bounds. In this work, we consider a Relative L2 sensitivity assumption, in which the bound on the distance between two query outputs may also depend on their norm. Leveraging this assumption, we introduce the Relative Gaussian Mechanism (RGM), in which the variance of the noise depends on the norm of the output. We prove tight bounds on the RDP parameters under relative L2 sensitivity, and characterize the privacy loss incurred by using output-dependent noise. In particular, we show that RGM naturally adapts to a latent variable that would control the norm of the output. Finally, we instantiate our framework to show tight guarantees for Private Gradient Descent, a problem that naturally fits our relative L2 sensitivity assumption.
format Preprint
id arxiv_https___arxiv_org_abs_2308_15250
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle The Relative Gaussian Mechanism and its Application to Private Gradient Descent
Hendrikx, Hadrien
Mangold, Paul
Bellet, Aurélien
Machine Learning
Cryptography and Security
Optimization and Control
The Gaussian Mechanism (GM), which consists in adding Gaussian noise to a vector-valued query before releasing it, is a standard privacy protection mechanism. In particular, given that the query respects some L2 sensitivity property (the L2 distance between outputs on any two neighboring inputs is bounded), GM guarantees Rényi Differential Privacy (RDP). Unfortunately, precisely bounding the L2 sensitivity can be hard, thus leading to loose privacy bounds. In this work, we consider a Relative L2 sensitivity assumption, in which the bound on the distance between two query outputs may also depend on their norm. Leveraging this assumption, we introduce the Relative Gaussian Mechanism (RGM), in which the variance of the noise depends on the norm of the output. We prove tight bounds on the RDP parameters under relative L2 sensitivity, and characterize the privacy loss incurred by using output-dependent noise. In particular, we show that RGM naturally adapts to a latent variable that would control the norm of the output. Finally, we instantiate our framework to show tight guarantees for Private Gradient Descent, a problem that naturally fits our relative L2 sensitivity assumption.
title The Relative Gaussian Mechanism and its Application to Private Gradient Descent
topic Machine Learning
Cryptography and Security
Optimization and Control
url https://arxiv.org/abs/2308.15250