GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks

Runzhe Hao; Rasheed Hussain; Juan Marcelo Parra-Ullauri; Xenofon Vasilakos; Reza Nejabati; Dimitra Simeonidou

doi:10.1109/INFOCOMWKSHPS61880.2024.10620772

GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks

Runzhe Hao, Rasheed Hussain, Juan Marcelo Parra-Ullauri, Xenofon Vasilakos, Reza Nejabati, Dimitra Simeonidou

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

4 Citations (Scopus)

Abstract

Federated Learning (FL) is vulnerable to various attacks including poisoning and inference. However, the existing offensive security evaluation of FL assumes that the attackers know data distribution. In this paper, we present a novel attack where FL participants carry out inference and privacy abuse attacks against the FL by leveraging Generating Adversarial Networks (GANs). The attacker (impersonating a benign participant) uses GAN to generate a similar dataset to other participants, and then covertly poisons the data. We demonstrated the attack successfully and tested it on two datasets, the IoT network traffic dataset and MNIST. The results reveal that for FL to be successfully used in IoT applications, protection against such attacks is critically essential.

Original language	English
Title of host publication	IEEE INFOCOM 2024 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
ISBN (Electronic)	9798350384475
ISBN (Print)	9798350384482
DOIs	https://doi.org/10.1109/INFOCOMWKSHPS61880.2024.10620772
Publication status	Published - 13 Aug 2024
Event	2024 IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2024 - Vancouver, Canada Duration: 20 May 2024 → 20 May 2024

Publication series

Name	IEEE Conference on Computer Communications Workshops
Publisher	IEEE
ISSN (Print)	2159-4228
ISSN (Electronic)	2833-0587

Conference

Conference	2024 IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2024
Country/Territory	Canada
City	Vancouver
Period	20/05/24 → 20/05/24

Bibliographical note

Publisher Copyright:
© 2024 IEEE.

Keywords

Federated Learning
GAN
IoT Security

Access to Document

10.1109/INFOCOMWKSHPS61880.2024.10620772

Handle.net

Persistent link

Cite this

Hao, R., Hussain, R., Parra-Ullauri, J. M., Vasilakos, X., Nejabati, R., & Simeonidou, D. (2024). GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks. In IEEE INFOCOM 2024 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS) (IEEE Conference on Computer Communications Workshops). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/INFOCOMWKSHPS61880.2024.10620772

@inproceedings{26240b71122c4d68acf5b0d3aaaf8d71,

title = "GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks",

abstract = "Federated Learning (FL) is vulnerable to various attacks including poisoning and inference. However, the existing offensive security evaluation of FL assumes that the attackers know data distribution. In this paper, we present a novel attack where FL participants carry out inference and privacy abuse attacks against the FL by leveraging Generating Adversarial Networks (GANs). The attacker (impersonating a benign participant) uses GAN to generate a similar dataset to other participants, and then covertly poisons the data. We demonstrated the attack successfully and tested it on two datasets, the IoT network traffic dataset and MNIST. The results reveal that for FL to be successfully used in IoT applications, protection against such attacks is critically essential.",

keywords = "Federated Learning, GAN, IoT Security",

author = "Runzhe Hao and Rasheed Hussain and Parra-Ullauri, \{Juan Marcelo\} and Xenofon Vasilakos and Reza Nejabati and Dimitra Simeonidou",

note = "Publisher Copyright: {\textcopyright} 2024 IEEE.; 2024 IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2024 ; Conference date: 20-05-2024 Through 20-05-2024",

year = "2024",

month = aug,

day = "13",

doi = "10.1109/INFOCOMWKSHPS61880.2024.10620772",

language = "English",

isbn = "9798350384482",

series = "IEEE Conference on Computer Communications Workshops",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

booktitle = "IEEE INFOCOM 2024 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)",

address = "United States",

}

Hao, R, Hussain, R, Parra-Ullauri, JM, Vasilakos, X, Nejabati, R & Simeonidou, D 2024, GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks. in IEEE INFOCOM 2024 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). IEEE Conference on Computer Communications Workshops, Institute of Electrical and Electronics Engineers (IEEE), 2024 IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2024, Vancouver, Canada, 20/05/24. https://doi.org/10.1109/INFOCOMWKSHPS61880.2024.10620772

GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks. / Hao, Runzhe; Hussain, Rasheed; Parra-Ullauri, Juan Marcelo et al.
IEEE INFOCOM 2024 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). Institute of Electrical and Electronics Engineers (IEEE), 2024. (IEEE Conference on Computer Communications Workshops).

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks

AU - Hao, Runzhe

AU - Hussain, Rasheed

AU - Parra-Ullauri, Juan Marcelo

AU - Vasilakos, Xenofon

AU - Nejabati, Reza

AU - Simeonidou, Dimitra

PY - 2024/8/13

Y1 - 2024/8/13

N2 - Federated Learning (FL) is vulnerable to various attacks including poisoning and inference. However, the existing offensive security evaluation of FL assumes that the attackers know data distribution. In this paper, we present a novel attack where FL participants carry out inference and privacy abuse attacks against the FL by leveraging Generating Adversarial Networks (GANs). The attacker (impersonating a benign participant) uses GAN to generate a similar dataset to other participants, and then covertly poisons the data. We demonstrated the attack successfully and tested it on two datasets, the IoT network traffic dataset and MNIST. The results reveal that for FL to be successfully used in IoT applications, protection against such attacks is critically essential.

AB - Federated Learning (FL) is vulnerable to various attacks including poisoning and inference. However, the existing offensive security evaluation of FL assumes that the attackers know data distribution. In this paper, we present a novel attack where FL participants carry out inference and privacy abuse attacks against the FL by leveraging Generating Adversarial Networks (GANs). The attacker (impersonating a benign participant) uses GAN to generate a similar dataset to other participants, and then covertly poisons the data. We demonstrated the attack successfully and tested it on two datasets, the IoT network traffic dataset and MNIST. The results reveal that for FL to be successfully used in IoT applications, protection against such attacks is critically essential.

KW - Federated Learning

KW - GAN

KW - IoT Security

UR - https://www.scopus.com/pages/publications/105003105320

U2 - 10.1109/INFOCOMWKSHPS61880.2024.10620772

DO - 10.1109/INFOCOMWKSHPS61880.2024.10620772

M3 - Conference Contribution (Conference Proceeding)

AN - SCOPUS:105003105320

SN - 9798350384482

T3 - IEEE Conference on Computer Communications Workshops

BT - IEEE INFOCOM 2024 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)

PB - Institute of Electrical and Electronics Engineers (IEEE)

T2 - 2024 IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2024

Y2 - 20 May 2024 through 20 May 2024

ER -

Hao R, Hussain R, Parra-Ullauri JM, Vasilakos X, Nejabati R, Simeonidou D. GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks. In IEEE INFOCOM 2024 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS). Institute of Electrical and Electronics Engineers (IEEE). 2024. (IEEE Conference on Computer Communications Workshops). doi: 10.1109/INFOCOMWKSHPS61880.2024.10620772

GAN-Based Privacy Abuse Attack on Federated Learning in IoT Networks

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Handle.net

Other files and links

Fingerprint

Cite this