Adaptive Federated Learning in Non-IID Data Environment

Jae Wook Lee; Haneul Ko

doi:10.7840/kics.2024.49.8.1118

Adaptive Federated Learning in Non-IID Data Environment

Jae Wook Lee, Haneul Ko

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

Abstract

In this paper, we propose the adaptive federated learning called FedA under non-IID data environment to guarantee the training accuracy and reduce training time. FedA adaptively selects the proper traditional federated learning scheme according to the non-IID degree. Also, we conduct the simulation to obtain the policy representing which federated learning scheme is configured according to the non-IID degree and to confirm the outperformance of our proposed scheme.

Original language	English
Pages (from-to)	1118-1120
Number of pages	3
Journal	Journal of Korean Institute of Communications and Information Sciences
Volume	49
Issue number	8
DOIs	https://doi.org/10.7840/kics.2024.49.8.1118
Publication status	Published - Aug 2024

Bibliographical note

Publisher Copyright:
© 2024, Korean Institute of Communications and Information Sciences. All rights reserved.

Keywords

Federated Learning
non-IID problem
training accuracy

Access to Document

10.7840/kics.2024.49.8.1118

Cite this

@article{f20cf3116ada45229a150063c1e13cb0,

title = "Adaptive Federated Learning in Non-IID Data Environment",

abstract = "In this paper, we propose the adaptive federated learning called FedA under non-IID data environment to guarantee the training accuracy and reduce training time. FedA adaptively selects the proper traditional federated learning scheme according to the non-IID degree. Also, we conduct the simulation to obtain the policy representing which federated learning scheme is configured according to the non-IID degree and to confirm the outperformance of our proposed scheme.",

keywords = "Federated Learning, non-IID problem, training accuracy",

author = "Lee, {Jae Wook} and Haneul Ko",

year = "2024",

month = aug,

doi = "10.7840/kics.2024.49.8.1118",

language = "English",

volume = "49",

pages = "1118--1120",

journal = "Journal of Korean Institute of Communications and Information Sciences",

issn = "1226-4717",

publisher = "Korean Institute of Communications and Information Sciences",

number = "8",

}

TY - JOUR

T1 - Adaptive Federated Learning in Non-IID Data Environment

AU - Lee, Jae Wook

AU - Ko, Haneul

PY - 2024/8

Y1 - 2024/8

N2 - In this paper, we propose the adaptive federated learning called FedA under non-IID data environment to guarantee the training accuracy and reduce training time. FedA adaptively selects the proper traditional federated learning scheme according to the non-IID degree. Also, we conduct the simulation to obtain the policy representing which federated learning scheme is configured according to the non-IID degree and to confirm the outperformance of our proposed scheme.

AB - In this paper, we propose the adaptive federated learning called FedA under non-IID data environment to guarantee the training accuracy and reduce training time. FedA adaptively selects the proper traditional federated learning scheme according to the non-IID degree. Also, we conduct the simulation to obtain the policy representing which federated learning scheme is configured according to the non-IID degree and to confirm the outperformance of our proposed scheme.

KW - Federated Learning

KW - non-IID problem

KW - training accuracy

UR - http://www.scopus.com/inward/record.url?scp=85207844877&partnerID=8YFLogxK

U2 - 10.7840/kics.2024.49.8.1118

DO - 10.7840/kics.2024.49.8.1118

M3 - Article

AN - SCOPUS:85207844877

SN - 1226-4717

VL - 49

SP - 1118

EP - 1120

JO - Journal of Korean Institute of Communications and Information Sciences

JF - Journal of Korean Institute of Communications and Information Sciences

IS - 8

ER -

Adaptive Federated Learning in Non-IID Data Environment

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this