Adaptive Deadline Determination for Mobile Device Selection in Federated Learning

Jaewook Lee; Haneul Ko; Sangheon Pack

doi:10.1109/TVT.2021.3136308

Adaptive Deadline Determination for Mobile Device Selection in Federated Learning

Jaewook Lee, Haneul Ko, Sangheon Pack

School of Electrical Engineering

Research output: Contribution to journal › Article › peer-review

6 Citations (Scopus)

Abstract

Owing to dynamically changing resources and channel conditions of mobile devices (MDs), when a static deadline-based MD selection scheme is used for federated learning, resource utilization of MDs can be degraded. To mitigate this problem, we propose an adaptive deadline determination (ADD) algorithm for MD selection, where a deadline for each round is adaptively determined with the consideration of the performance disparity of MDs. Evaluation results demonstrate that ADD can achieve the fastest average convergence time among the comparison schemes.

Original language	English
Pages (from-to)	3367-3371
Number of pages	5
Journal	IEEE Transactions on Vehicular Technology
Volume	71
Issue number	3
DOIs	https://doi.org/10.1109/TVT.2021.3136308
Publication status	Published - 2022 Mar 1

Keywords

Federated learning
adaptive deadline
mobile device selection

ASJC Scopus subject areas

Automotive Engineering
Aerospace Engineering
Electrical and Electronic Engineering
Applied Mathematics

Access to Document

10.1109/TVT.2021.3136308

Cite this

@article{5f92f68dad144290a5a4168301805292,

title = "Adaptive Deadline Determination for Mobile Device Selection in Federated Learning",

abstract = "Owing to dynamically changing resources and channel conditions of mobile devices (MDs), when a static deadline-based MD selection scheme is used for federated learning, resource utilization of MDs can be degraded. To mitigate this problem, we propose an adaptive deadline determination (ADD) algorithm for MD selection, where a deadline for each round is adaptively determined with the consideration of the performance disparity of MDs. Evaluation results demonstrate that ADD can achieve the fastest average convergence time among the comparison schemes.",

keywords = "Federated learning, adaptive deadline, mobile device selection",

author = "Jaewook Lee and Haneul Ko and Sangheon Pack",

note = "Funding Information: This work was supported by the Future Combat System Network Technology Research Center program of Defense Acquisition Program Administration and Agency for Defense Development under Grant UD190033ED. The review of this article was coordinated by Dr. Ai-Chun Pang. Publisher Copyright: {\textcopyright} 1967-2012 IEEE.",

year = "2022",

month = mar,

day = "1",

doi = "10.1109/TVT.2021.3136308",

language = "English",

volume = "71",

pages = "3367--3371",

journal = "IEEE Transactions on Vehicular Technology",

issn = "0018-9545",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "3",

}

TY - JOUR

T1 - Adaptive Deadline Determination for Mobile Device Selection in Federated Learning

AU - Lee, Jaewook

AU - Ko, Haneul

AU - Pack, Sangheon

N1 - Funding Information: This work was supported by the Future Combat System Network Technology Research Center program of Defense Acquisition Program Administration and Agency for Defense Development under Grant UD190033ED. The review of this article was coordinated by Dr. Ai-Chun Pang. Publisher Copyright: © 1967-2012 IEEE.

PY - 2022/3/1

Y1 - 2022/3/1

N2 - Owing to dynamically changing resources and channel conditions of mobile devices (MDs), when a static deadline-based MD selection scheme is used for federated learning, resource utilization of MDs can be degraded. To mitigate this problem, we propose an adaptive deadline determination (ADD) algorithm for MD selection, where a deadline for each round is adaptively determined with the consideration of the performance disparity of MDs. Evaluation results demonstrate that ADD can achieve the fastest average convergence time among the comparison schemes.

AB - Owing to dynamically changing resources and channel conditions of mobile devices (MDs), when a static deadline-based MD selection scheme is used for federated learning, resource utilization of MDs can be degraded. To mitigate this problem, we propose an adaptive deadline determination (ADD) algorithm for MD selection, where a deadline for each round is adaptively determined with the consideration of the performance disparity of MDs. Evaluation results demonstrate that ADD can achieve the fastest average convergence time among the comparison schemes.

KW - Federated learning

KW - adaptive deadline

KW - mobile device selection

UR - http://www.scopus.com/inward/record.url?scp=85121803344&partnerID=8YFLogxK

U2 - 10.1109/TVT.2021.3136308

DO - 10.1109/TVT.2021.3136308

M3 - Article

AN - SCOPUS:85121803344

SN - 0018-9545

VL - 71

SP - 3367

EP - 3371

JO - IEEE Transactions on Vehicular Technology

JF - IEEE Transactions on Vehicular Technology

IS - 3

ER -

Adaptive Deadline Determination for Mobile Device Selection in Federated Learning

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this