Cross-silo Federated Learning with Record-level Personalized Differential Privacy

Federated learning (FL) enhanced by differential privacy has emerged as a popular approach to better safeguard the privacy of client-side data by protecting clients' contributions during the training process. Existing solutions typically assume a uniform privacy budget for all records and provi...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Proceedings of the ... ACM Conference on Computer and Communications Security Ročník 2024; s. 303
Hlavní autoři: Liu, Junxu, Lou, Jian, Xiong, Li, Liu, Jinfei, Meng, Xiaofeng
Médium: Journal Article
Jazyk:angličtina
Vydáno: United States 01.10.2024
Témata:
Differential Privacy
Federated Learning
Personalized Privacy Protection
ISSN:1543-7221, 1543-7221
On-line přístup:Zjistit podrobnosti o přístupu
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Federated learning (FL) enhanced by differential privacy has emerged as a popular approach to better safeguard the privacy of client-side data by protecting clients' contributions during the training process. Existing solutions typically assume a uniform privacy budget for all records and provide one-size-fits-all solutions that may not be adequate to meet each record's privacy requirement. In this paper, we explore the uncharted territory of cross-silo FL with record-level personalized differential privacy. We devise a novel framework named , employing a two-stage hybrid sampling scheme with both uniform client-level sampling and non-uniform record-level sampling to accommodate varying privacy requirements. A critical and non-trivial problem is how to determine the ideal per-record sampling probability given the personalized privacy budget . We introduce a versatile solution named , allowing us to uncover a significant insight into the nonlinear correlation between and and derive an elegant mathematical model to tackle the problem. Our evaluation demonstrates that our solution can provide significant performance gains over the baselines that do not consider personalized privacy preservation.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:1543-7221
1543-7221
DOI:10.1145/3658644.3670351