Guardado en:
Detalles Bibliográficos
Autores principales: Slokom, Manel, de Wolf, Peter-Paul, Larson, Martha
Formato: Preprint
Publicado: 2023
Materias:
Acceso en línea:https://arxiv.org/abs/2310.08775
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910451941507072
author Slokom, Manel
de Wolf, Peter-Paul
Larson, Martha
author_facet Slokom, Manel
de Wolf, Peter-Paul
Larson, Martha
contents We investigate an attack on a machine learning model that predicts whether a person or household will relocate in the next two years, i.e., a propensity-to-move classifier. The attack assumes that the attacker can query the model to obtain predictions and that the marginal distribution of the data on which the model was trained is publicly available. The attack also assumes that the attacker has obtained the values of non-sensitive attributes for a certain number of target individuals. The objective of the attack is to infer the values of sensitive attributes for these target individuals. We explore how replacing the original data with synthetic data when training the model impacts how successfully the attacker can infer sensitive attributes.
format Preprint
id arxiv_https___arxiv_org_abs_2310_08775
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle When Machine Learning Models Leak: An Exploration of Synthetic Training Data
Slokom, Manel
de Wolf, Peter-Paul
Larson, Martha
Machine Learning
We investigate an attack on a machine learning model that predicts whether a person or household will relocate in the next two years, i.e., a propensity-to-move classifier. The attack assumes that the attacker can query the model to obtain predictions and that the marginal distribution of the data on which the model was trained is publicly available. The attack also assumes that the attacker has obtained the values of non-sensitive attributes for a certain number of target individuals. The objective of the attack is to infer the values of sensitive attributes for these target individuals. We explore how replacing the original data with synthetic data when training the model impacts how successfully the attacker can infer sensitive attributes.
title When Machine Learning Models Leak: An Exploration of Synthetic Training Data
topic Machine Learning
url https://arxiv.org/abs/2310.08775