Cross-silo Federated Learning with Record-level Personalized Differential Privacy
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909233539186688 |
|---|---|
| author | Liu, Junxu Lou, Jian Xiong, Li Liu, Jinfei Meng, Xiaofeng |
| author_facet | Liu, Junxu Lou, Jian Xiong, Li Liu, Jinfei Meng, Xiaofeng |
| contents | 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 \textit{rPDP-FL}, 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 $q$ given the personalized privacy budget $\varepsilon$. We introduce a versatile solution named \textit{Simulation-CurveFitting}, allowing us to uncover a significant insight into the nonlinear correlation between $q$ and $\varepsilon$ 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_16251 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Cross-silo Federated Learning with Record-level Personalized Differential Privacy Liu, Junxu Lou, Jian Xiong, Li Liu, Jinfei Meng, Xiaofeng Cryptography and Security Artificial Intelligence Machine Learning 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 \textit{rPDP-FL}, 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 $q$ given the personalized privacy budget $\varepsilon$. We introduce a versatile solution named \textit{Simulation-CurveFitting}, allowing us to uncover a significant insight into the nonlinear correlation between $q$ and $\varepsilon$ 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. |
| title | Cross-silo Federated Learning with Record-level Personalized Differential Privacy |
| topic | Cryptography and Security Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2401.16251 |