Home > Research > Publications & Outputs > Privacy-preserving Decentralized Federated Lear...

Research

Associated organisational units

Links

Text available via DOI:

View graph of relations

Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Published

Standard

Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph. / Lu, Yang; Yu, Zhengxin; Suri, Neeraj.
In: arXiv, Vol. abs/2210.00325, 01.10.2022.

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Harvard

Lu, Y, Yu, Z & Suri, N 2022, 'Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph', arXiv, vol. abs/2210.00325. https://doi.org/10.48550/arXiv.2210.00325

APA

Lu, Y., Yu, Z., & Suri, N. (2022). Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph. arXiv, abs/2210.00325. https://doi.org/10.48550/arXiv.2210.00325

Vancouver

Lu Y, Yu Z, Suri N. Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph. arXiv. 2022 Oct 1;abs/2210.00325. doi: 10.48550/arXiv.2210.00325

Author

Lu, Yang ; Yu, Zhengxin ; Suri, Neeraj. / Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph. In: arXiv. 2022 ; Vol. abs/2210.00325.

Bibtex

@article{c4d1a85221b34183bbcea88cceaaa309,
title = "Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph",
abstract = "Establishing how a set of learners can provide privacy-preserving federated learning in a fully decentralized (peer-to-peer, no coordinator) manner is an open problem. We propose the first privacy-preserving consensus-based algorithm for the distributed learners to achieve decentralized global model aggregation in an environment of high mobility, where the communication graph between the learners may vary between successive rounds of model aggregation. In particular, in each round of global model aggregation, the Metropolis-Hastings method is applied to update the weighted adjacency matrix based on the current communication topology. In addition, the Shamir's secret sharing scheme is integrated to facilitate privacy in reaching consensus of the global model. The paper establishes the correctness and privacy properties of the proposed algorithm. The computational efficiency is evaluated by a simulation built on a federated learning framework with a real-word dataset.",
author = "Yang Lu and Zhengxin Yu and Neeraj Suri",
year = "2022",
month = oct,
day = "1",
doi = "10.48550/arXiv.2210.00325",
language = "English",
volume = "abs/2210.00325",
journal = "arXiv",
issn = "2331-8422",

}

RIS

TY - JOUR

T1 - Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph

AU - Lu, Yang

AU - Yu, Zhengxin

AU - Suri, Neeraj

PY - 2022/10/1

Y1 - 2022/10/1

N2 - Establishing how a set of learners can provide privacy-preserving federated learning in a fully decentralized (peer-to-peer, no coordinator) manner is an open problem. We propose the first privacy-preserving consensus-based algorithm for the distributed learners to achieve decentralized global model aggregation in an environment of high mobility, where the communication graph between the learners may vary between successive rounds of model aggregation. In particular, in each round of global model aggregation, the Metropolis-Hastings method is applied to update the weighted adjacency matrix based on the current communication topology. In addition, the Shamir's secret sharing scheme is integrated to facilitate privacy in reaching consensus of the global model. The paper establishes the correctness and privacy properties of the proposed algorithm. The computational efficiency is evaluated by a simulation built on a federated learning framework with a real-word dataset.

AB - Establishing how a set of learners can provide privacy-preserving federated learning in a fully decentralized (peer-to-peer, no coordinator) manner is an open problem. We propose the first privacy-preserving consensus-based algorithm for the distributed learners to achieve decentralized global model aggregation in an environment of high mobility, where the communication graph between the learners may vary between successive rounds of model aggregation. In particular, in each round of global model aggregation, the Metropolis-Hastings method is applied to update the weighted adjacency matrix based on the current communication topology. In addition, the Shamir's secret sharing scheme is integrated to facilitate privacy in reaching consensus of the global model. The paper establishes the correctness and privacy properties of the proposed algorithm. The computational efficiency is evaluated by a simulation built on a federated learning framework with a real-word dataset.

U2 - 10.48550/arXiv.2210.00325

DO - 10.48550/arXiv.2210.00325

M3 - Journal article

VL - abs/2210.00325

JO - arXiv

JF - arXiv

SN - 2331-8422

ER -