File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Edge Federated Learning via Unit-Modulus Over-The-Air Computation

TitleEdge Federated Learning via Unit-Modulus Over-The-Air Computation
Authors
KeywordsAutonomous driving
federated learning
large-scale optimization
over-the-air computation
Issue Date1-May-2022
PublisherInstitute of Electrical and Electronics Engineers
Citation
IEEE Transactions on Communications, 2022, v. 70, n. 5, p. 3141-3156 How to Cite?
Abstract

Edge federated learning (FL) is an emerging paradigm that trains a global parametric model from distributed datasets based on wireless communications. This paper proposes a unit-modulus over-the-air computation (UMAirComp) framework to facilitate efficient edge federated learning, which simultaneously uploads local model parameters and updates global model parameters via analog beamforming. The proposed framework avoids sophisticated baseband signal processing, leading to low communication delays and implementation costs. Training loss bounds of UMAirComp FL systems are derived and two low-complexity large-scale optimization algorithms, termed penalty alternating minimization (PAM) and accelerated gradient projection (AGP), are proposed to minimize the nonconvex nonsmooth loss bound. Simulation results show that the proposed UMAirComp framework with PAM algorithm achieves a smaller mean square error of model parameters' estimation, training loss, and test error compared with other benchmark schemes. Moreover, the proposed UMAirComp framework with AGP algorithm achieves satisfactory performance while reduces the computational complexity by orders of magnitude compared with existing optimization algorithms. Finally, we demonstrate the implementation of UMAirComp in a vehicle-to-everything autonomous driving simulation platform. It is found that autonomous driving tasks are more sensitive to model parameter errors than other tasks since the neural networks for autonomous driving contain sparser model parameters.


Persistent Identifierhttp://hdl.handle.net/10722/339295
ISSN
2021 Impact Factor: 6.166
2020 SCImago Journal Rankings: 1.468
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorWang, Shuai-
dc.contributor.authorHong, Yuncong-
dc.contributor.authorWang, Rui-
dc.contributor.authorHao, Qi-
dc.contributor.authorWu, Yik-Chung-
dc.contributor.authorNg, Derrick Wing Kwan -
dc.date.accessioned2024-03-11T10:35:29Z-
dc.date.available2024-03-11T10:35:29Z-
dc.date.issued2022-05-01-
dc.identifier.citationIEEE Transactions on Communications, 2022, v. 70, n. 5, p. 3141-3156-
dc.identifier.issn0090-6778-
dc.identifier.urihttp://hdl.handle.net/10722/339295-
dc.description.abstract<p>Edge federated learning (FL) is an emerging paradigm that trains a global parametric model from distributed datasets based on wireless communications. This paper proposes a unit-modulus over-the-air computation (UMAirComp) framework to facilitate efficient edge federated learning, which simultaneously uploads local model parameters and updates global model parameters via analog beamforming. The proposed framework avoids sophisticated baseband signal processing, leading to low communication delays and implementation costs. Training loss bounds of UMAirComp FL systems are derived and two low-complexity large-scale optimization algorithms, termed penalty alternating minimization (PAM) and accelerated gradient projection (AGP), are proposed to minimize the nonconvex nonsmooth loss bound. Simulation results show that the proposed UMAirComp framework with PAM algorithm achieves a smaller mean square error of model parameters' estimation, training loss, and test error compared with other benchmark schemes. Moreover, the proposed UMAirComp framework with AGP algorithm achieves satisfactory performance while reduces the computational complexity by orders of magnitude compared with existing optimization algorithms. Finally, we demonstrate the implementation of UMAirComp in a vehicle-to-everything autonomous driving simulation platform. It is found that autonomous driving tasks are more sensitive to model parameter errors than other tasks since the neural networks for autonomous driving contain sparser model parameters.<br></p>-
dc.languageeng-
dc.publisherInstitute of Electrical and Electronics Engineers-
dc.relation.ispartofIEEE Transactions on Communications-
dc.rightsThis work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.-
dc.subjectAutonomous driving-
dc.subjectfederated learning-
dc.subjectlarge-scale optimization-
dc.subjectover-the-air computation-
dc.titleEdge Federated Learning via Unit-Modulus Over-The-Air Computation-
dc.typeArticle-
dc.identifier.doi10.1109/TCOMM.2022.3153488-
dc.identifier.scopuseid_2-s2.0-85125359212-
dc.identifier.volume70-
dc.identifier.issue5-
dc.identifier.spage3141-
dc.identifier.epage3156-
dc.identifier.eissn1558-0857-
dc.identifier.isiWOS:000797439600022-
dc.identifier.issnl0090-6778-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats