(Local) Differential Privacy has NO Disparate Impact on Fairness - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

(Local) Differential Privacy has NO Disparate Impact on Fairness

Résumé

In recent years, Local Differential Privacy (LDP), a robust privacy-preserving methodology, has gained widespread adoption in realworld applications. With LDP, users can perturb their data on their devices before sending it out for analysis. However, as the collection of multiple sensitive information becomes more prevalent across various industries, collecting a single sensitive attribute under LDP may not be sufficient. Correlated attributes in the data may still lead to inferences about the sensitive attribute. This paper empirically studies the impact of collecting multiple sensitive attributes under LDP on fairness. We propose a novel privacy budget allocation scheme that considers the varying domain size of sensitive attributes. This generally led to a better privacyutility-fairness trade-off in our experiments than the state-of-art solution. Our results show that LDP leads to slightly improved fairness in learning problems without significantly affecting the performance of the models. We conduct extensive experiments evaluating three benchmark datasets using several group fairness metrics and seven state-of-the-art LDP protocols. Overall, this study challenges the common belief that differential privacy necessarily leads to worsened fairness in machine learning.
Fichier principal
Vignette du fichier
2023_Impact_Fairness_LDP.pdf (931.09 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Licence : CC BY - Paternité

Dates et versions

hal-04175027 , version 1 (01-08-2023)

Licence

Paternité

Identifiants

Citer

Héber Hwang Arcolezi, Karima Makhlouf, Catuscia Palamidessi. (Local) Differential Privacy has NO Disparate Impact on Fairness. DBSec 2023 - 37th IFIP Annual Conference on Data and Applications Security and Privacy, Vijay Atluri; Anna Lisa Ferrara, Jul 2023, SOPHIA ANTIPOLIS, France. pp.3-21, ⟨10.1007/978-3-031-37586-6_1⟩. ⟨hal-04175027⟩
61 Consultations
50 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More