An Architecture for Resilient Federated Learning Through Parallel Recognition

Jeongeun Kim; Youngwoo Jeong; Suyeon Jang; Seung Eun Lee

doi:10.1145/3559009.3569689

An Architecture for Resilient Federated Learning Through Parallel Recognition

Jeongeun Kim, Youngwoo Jeong, Suyeon Jang, Seung Eun Lee

Electronic Engineering Program

Seoul National University of Science and Technology (SNUST)

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

In federated learning, non-independent and identically distributed (non-IID) local datasets lead to accuracy loss compared to homogeneous distribution of datasets. In this paper, we propose an architecture for improving accuracy and offering resilience through federation utilizing non-IID datasets. The proposed architecture performs parallel recognition employing triplication of AI processors with different intelligence. Experimental results demonstrate that the proposed architecture improves accuracy by 2.3% compared to accuracy of a single AI processor on average and guarantees resilience.

Original language	English
Title of host publication	PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	546-547
Number of pages	2
ISBN (Electronic)	9781450398688
DOIs	https://doi.org/10.1145/3559009.3569689
State	Published - 8 Oct 2022
Event	31st International Conference on Parallel Architectures and Compilation Techniques, PACT 2022 - Chicago, United States Duration: 8 Oct 2022 → 10 Oct 2022

Publication series

Name	Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT
ISSN (Print)	1089-795X

Conference

Conference	31st International Conference on Parallel Architectures and Compilation Techniques, PACT 2022
Country/Territory	United States
City	Chicago
Period	8/10/22 → 10/10/22

Access to Document

10.1145/3559009.3569689

Cite this

Kim, J., Jeong, Y., Jang, S., & Lee, S. E. (2022). An Architecture for Resilient Federated Learning Through Parallel Recognition. In PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques (pp. 546-547). (Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1145/3559009.3569689

Kim, Jeongeun ; Jeong, Youngwoo ; Jang, Suyeon et al. / An Architecture for Resilient Federated Learning Through Parallel Recognition. PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques. Institute of Electrical and Electronics Engineers Inc., 2022. pp. 546-547 (Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT).

@inproceedings{7f37d2d12048459f996b4c2706246d52,

title = "An Architecture for Resilient Federated Learning Through Parallel Recognition",

abstract = "In federated learning, non-independent and identically distributed (non-IID) local datasets lead to accuracy loss compared to homogeneous distribution of datasets. In this paper, we propose an architecture for improving accuracy and offering resilience through federation utilizing non-IID datasets. The proposed architecture performs parallel recognition employing triplication of AI processors with different intelligence. Experimental results demonstrate that the proposed architecture improves accuracy by 2.3\% compared to accuracy of a single AI processor on average and guarantees resilience.",

author = "Jeongeun Kim and Youngwoo Jeong and Suyeon Jang and Lee, \{Seung Eun\}",

note = "Publisher Copyright: {\textcopyright} 2022 Association for Computing Machinery.; 31st International Conference on Parallel Architectures and Compilation Techniques, PACT 2022 ; Conference date: 08-10-2022 Through 10-10-2022",

year = "2022",

month = oct,

day = "8",

doi = "10.1145/3559009.3569689",

language = "English",

series = "Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT",

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

pages = "546--547",

booktitle = "PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques",

}

Kim, J, Jeong, Y, Jang, S & Lee, SE 2022, An Architecture for Resilient Federated Learning Through Parallel Recognition. in PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques. Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT, Institute of Electrical and Electronics Engineers Inc., pp. 546-547, 31st International Conference on Parallel Architectures and Compilation Techniques, PACT 2022, Chicago, United States, 8/10/22. https://doi.org/10.1145/3559009.3569689

An Architecture for Resilient Federated Learning Through Parallel Recognition. / Kim, Jeongeun; Jeong, Youngwoo; Jang, Suyeon et al.
PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques. Institute of Electrical and Electronics Engineers Inc., 2022. p. 546-547 (Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - An Architecture for Resilient Federated Learning Through Parallel Recognition

AU - Kim, Jeongeun

AU - Jeong, Youngwoo

AU - Jang, Suyeon

AU - Lee, Seung Eun

PY - 2022/10/8

Y1 - 2022/10/8

N2 - In federated learning, non-independent and identically distributed (non-IID) local datasets lead to accuracy loss compared to homogeneous distribution of datasets. In this paper, we propose an architecture for improving accuracy and offering resilience through federation utilizing non-IID datasets. The proposed architecture performs parallel recognition employing triplication of AI processors with different intelligence. Experimental results demonstrate that the proposed architecture improves accuracy by 2.3% compared to accuracy of a single AI processor on average and guarantees resilience.

AB - In federated learning, non-independent and identically distributed (non-IID) local datasets lead to accuracy loss compared to homogeneous distribution of datasets. In this paper, we propose an architecture for improving accuracy and offering resilience through federation utilizing non-IID datasets. The proposed architecture performs parallel recognition employing triplication of AI processors with different intelligence. Experimental results demonstrate that the proposed architecture improves accuracy by 2.3% compared to accuracy of a single AI processor on average and guarantees resilience.

UR - https://www.scopus.com/pages/publications/85147331470

U2 - 10.1145/3559009.3569689

DO - 10.1145/3559009.3569689

M3 - Conference contribution

AN - SCOPUS:85147331470

T3 - Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT

SP - 546

EP - 547

BT - PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 31st International Conference on Parallel Architectures and Compilation Techniques, PACT 2022

Y2 - 8 October 2022 through 10 October 2022

ER -

Kim J, Jeong Y, Jang S, Lee SE. An Architecture for Resilient Federated Learning Through Parallel Recognition. In PACT 2022 - Proceedings of the 2022 International Conference on Parallel Architectures and Compilation Techniques. Institute of Electrical and Electronics Engineers Inc. 2022. p. 546-547. (Parallel Architectures and Compilation Techniques - Conference Proceedings, PACT). doi: 10.1145/3559009.3569689

An Architecture for Resilient Federated Learning Through Parallel Recognition

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this