PreprintCC BY 4.0arXiv
A Simple Presentation of Robust Membership Inference Attacks
Abstract
In a membership inference attack, an attacker is given the data of a particular individual, and tries to determine if that individual was included in a particular dataset, given some kind of statistics derived from the dataset or a model trained on the dataset. Roughly, a membership inference attack is robust if it succeeds even when the statistics are highly noisy, regardless of how that noise is introduced. We give a simple presentation of robust membership inference attacks for mean estimation, covering both multivariate Gaussian distributions and binary product distributions. We also give a summary of the literature on robust membership inference, covering history, extensions, and connections to other research areas.
Related
- Saturation in G: simple & robust causal inference in cluster randomized trials with informative cluster sizes
- Robust Direct Preference Optimization under Preference Label-Flip Attacks
- Black-Box Membership Inference via Word-Level Probability Estimation
- Membership Inference via Pairwise Likelihood Ratios
- Byzantine-Robust Federated RAG via Aligned Calibration and Fixed-Membership Conformal Prediction
BibTeX
@misc{gaucher2026simple,
title = {{A Simple Presentation of Robust Membership Inference Attacks}},
author = {Renaud Gaucher and Ephraim Linder and Adam Smith and Jonathan Ullman and Lydia Zakynthinou},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.36458},
eprint = {2610.36458},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.36458},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A Simple Presentation of Robust Membership Inference Attacks
AU - Renaud Gaucher
AU - Ephraim Linder
AU - Adam Smith
AU - Jonathan Ullman
AU - Lydia Zakynthinou
PY - 2026
DA - 2026/10/27
PB - arXiv
DO - 10.48550/arxiv.2610.36458
UR - https://arxiv.org/abs/2610.36458
AB - In a membership inference attack, an attacker is given the data of a particular individual, and tries to determine if that individual was included in a particular dataset, given some kind of statistics derived from the dataset or a model trained on the dataset. Roughly, a membership inference attack is robust if it succeeds even when the statistics are highly noisy, regardless of how that noise is introduced. We give a simple presentation of robust membership inference attacks for mean estimation, covering both multivariate Gaussian distributions and binary product distributions. We also give a summary of the literature on robust membership inference, covering history, extensions, and connections to other research areas.
ER -