Privacy Vulnerabilities in Marginals-based Synthetic Data

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Golob, Steven, Pentyala, Sikha, Maratkhan, Anuar, De Cock, Martine
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908292256628736
author Golob, Steven
Pentyala, Sikha
Maratkhan, Anuar
De Cock, Martine
author_facet Golob, Steven
Pentyala, Sikha
Maratkhan, Anuar
De Cock, Martine
contents When acting as a privacy-enhancing technology, synthetic data generation (SDG) aims to maintain a resemblance to the real data while excluding personally-identifiable information. Many SDG algorithms provide robust differential privacy (DP) guarantees to this end. However, we show that the strongest class of SDG algorithms--those that preserve \textit{marginal probabilities}, or similar statistics, from the underlying data--leak information about individuals that can be recovered more efficiently than previously understood. We demonstrate this by presenting a novel membership inference attack, MAMA-MIA, and evaluate it against three seminal DP SDG algorithms: MST, PrivBayes, and Private-GSD. MAMA-MIA leverages knowledge of which SDG algorithm was used, allowing it to learn information about the hidden data more accurately, and orders-of-magnitude faster, than other leading attacks. We use MAMA-MIA to lend insight into existing SDG vulnerabilities. Our approach went on to win the first SNAKE (SaNitization Algorithm under attacK ... $\varepsilon$) competition.
format Preprint
id arxiv_https___arxiv_org_abs_2410_05506
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Privacy Vulnerabilities in Marginals-based Synthetic Data
Golob, Steven
Pentyala, Sikha
Maratkhan, Anuar
De Cock, Martine
Cryptography and Security
Machine Learning
When acting as a privacy-enhancing technology, synthetic data generation (SDG) aims to maintain a resemblance to the real data while excluding personally-identifiable information. Many SDG algorithms provide robust differential privacy (DP) guarantees to this end. However, we show that the strongest class of SDG algorithms--those that preserve \textit{marginal probabilities}, or similar statistics, from the underlying data--leak information about individuals that can be recovered more efficiently than previously understood. We demonstrate this by presenting a novel membership inference attack, MAMA-MIA, and evaluate it against three seminal DP SDG algorithms: MST, PrivBayes, and Private-GSD. MAMA-MIA leverages knowledge of which SDG algorithm was used, allowing it to learn information about the hidden data more accurately, and orders-of-magnitude faster, than other leading attacks. We use MAMA-MIA to lend insight into existing SDG vulnerabilities. Our approach went on to win the first SNAKE (SaNitization Algorithm under attacK ... $\varepsilon$) competition.
title Privacy Vulnerabilities in Marginals-based Synthetic Data
topic Cryptography and Security
Machine Learning
url https://arxiv.org/abs/2410.05506