Boosting Communication Efficiency of Federated Learning's Secure Aggregation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Nazemi, Niousha, Tavallaie, Omid, Chen, Shuaijun, Zomaya, Albert Y., Holz, Ralph
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909187984850944
author Nazemi, Niousha
Tavallaie, Omid
Chen, Shuaijun
Zomaya, Albert Y.
Holz, Ralph
author_facet Nazemi, Niousha
Tavallaie, Omid
Chen, Shuaijun
Zomaya, Albert Y.
Holz, Ralph
contents Federated Learning (FL) is a decentralized machine learning approach where client devices train models locally and send them to a server that performs aggregation to generate a global model. FL is vulnerable to model inversion attacks, where the server can infer sensitive client data from trained models. Google's Secure Aggregation (SecAgg) protocol addresses this data privacy issue by masking each client's trained model using shared secrets and individual elements generated locally on the client's device. Although SecAgg effectively preserves privacy, it imposes considerable communication and computation overhead, especially as network size increases. Building upon SecAgg, this poster introduces a Communication-Efficient Secure Aggregation (CESA) protocol that substantially reduces this overhead by using only two shared secrets per client to mask the model. We propose our method for stable networks with low delay variation and limited client dropouts. CESA is independent of the data distribution and network size (for higher than 6 nodes), preventing the honest-but-curious server from accessing unmasked models. Our initial evaluation reveals that CESA significantly reduces the communication cost compared to SecAgg.
format Preprint
id arxiv_https___arxiv_org_abs_2405_01144
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Boosting Communication Efficiency of Federated Learning's Secure Aggregation
Nazemi, Niousha
Tavallaie, Omid
Chen, Shuaijun
Zomaya, Albert Y.
Holz, Ralph
Cryptography and Security
Federated Learning (FL) is a decentralized machine learning approach where client devices train models locally and send them to a server that performs aggregation to generate a global model. FL is vulnerable to model inversion attacks, where the server can infer sensitive client data from trained models. Google's Secure Aggregation (SecAgg) protocol addresses this data privacy issue by masking each client's trained model using shared secrets and individual elements generated locally on the client's device. Although SecAgg effectively preserves privacy, it imposes considerable communication and computation overhead, especially as network size increases. Building upon SecAgg, this poster introduces a Communication-Efficient Secure Aggregation (CESA) protocol that substantially reduces this overhead by using only two shared secrets per client to mask the model. We propose our method for stable networks with low delay variation and limited client dropouts. CESA is independent of the data distribution and network size (for higher than 6 nodes), preventing the honest-but-curious server from accessing unmasked models. Our initial evaluation reveals that CESA significantly reduces the communication cost compared to SecAgg.
title Boosting Communication Efficiency of Federated Learning's Secure Aggregation
topic Cryptography and Security
url https://arxiv.org/abs/2405.01144