ALI-DPFL: Differentially Private Federated Learning with Adaptive Local Iterations

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ling, Xinpeng, Fu, Jie, Wang, Kuncan, Liu, Haitao, Chen, Zhili
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917673819963392
author Ling, Xinpeng
Fu, Jie
Wang, Kuncan
Liu, Haitao
Chen, Zhili
author_facet Ling, Xinpeng
Fu, Jie
Wang, Kuncan
Liu, Haitao
Chen, Zhili
contents Federated Learning (FL) is a distributed machine learning technique that allows model training among multiple devices or organizations by sharing training parameters instead of raw data. However, adversaries can still infer individual information through inference attacks (e.g. differential attacks) on these training parameters. As a result, Differential Privacy (DP) has been widely used in FL to prevent such attacks. We consider differentially private federated learning in a resource-constrained scenario, where both privacy budget and communication rounds are constrained. By theoretically analyzing the convergence, we can find the optimal number of local DPSGD iterations for clients between any two sequential global updates. Based on this, we design an algorithm of Differentially Private Federated Learning with Adaptive Local Iterations (ALI-DPFL). We experiment our algorithm on the MNIST, FashionMNIST and Cifar10 datasets, and demonstrate significantly better performances than previous work in the resource-constraint scenario. Code is available at https://github.com/cheng-t/ALI-DPFL.
format Preprint
id arxiv_https___arxiv_org_abs_2308_10457
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle ALI-DPFL: Differentially Private Federated Learning with Adaptive Local Iterations
Ling, Xinpeng
Fu, Jie
Wang, Kuncan
Liu, Haitao
Chen, Zhili
Machine Learning
Cryptography and Security
Federated Learning (FL) is a distributed machine learning technique that allows model training among multiple devices or organizations by sharing training parameters instead of raw data. However, adversaries can still infer individual information through inference attacks (e.g. differential attacks) on these training parameters. As a result, Differential Privacy (DP) has been widely used in FL to prevent such attacks. We consider differentially private federated learning in a resource-constrained scenario, where both privacy budget and communication rounds are constrained. By theoretically analyzing the convergence, we can find the optimal number of local DPSGD iterations for clients between any two sequential global updates. Based on this, we design an algorithm of Differentially Private Federated Learning with Adaptive Local Iterations (ALI-DPFL). We experiment our algorithm on the MNIST, FashionMNIST and Cifar10 datasets, and demonstrate significantly better performances than previous work in the resource-constraint scenario. Code is available at https://github.com/cheng-t/ALI-DPFL.
title ALI-DPFL: Differentially Private Federated Learning with Adaptive Local Iterations
topic Machine Learning
Cryptography and Security
url https://arxiv.org/abs/2308.10457