Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping

Amit Singha; Ziqian Bi; Tao Li; Yimin Chen; Yanchao Zhang

doi:10.1145/3643833.3656123

Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping

Amit Singha, Ziqian Bi, Tao Li, Yimin Chen, Yanchao Zhang

Engineering, Ira A. Fulton Schools of (IAFSE)

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

Abstract

Wireless Human Activity Recognition (HAR), leveraging their non-intrusive nature, has the potential to revolutionize various sectors, including healthcare, virtual reality, and surveillance. The advent of millimeter wave (mmWave) technology has significantly enhanced the capabilities of wireless HAR systems. This paper presents the first systematic study on the vulnerabilities of mmWave-based HAR to label flipping poisoning attacks in the context of supervised contrastive learning. We identify three label poisoning attacks on the contrastive mmWave-based HAR and propose corresponding countermeasures. The efficacy of the attacks and also our countermeasures are experimentally validated on a prototype system. The attacks and countermeasures can be easily extended to other wireless HAR systems, thereby promoting security considerations in system design and deployment.

Original language	English (US)
Title of host publication	WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks
Publisher	Association for Computing Machinery, Inc
Pages	31-41
Number of pages	11
ISBN (Electronic)	9798400705823
DOIs	https://doi.org/10.1145/3643833.3656123
State	Published - May 27 2024
Event	17th ACM Conference on Security and Privacy in Wireless and Mobile Networks, WiSec 2024 - Seoul, Korea, Republic of Duration: May 27 2024 → May 29 2024

Publication series

Name	WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks

Conference

Conference	17th ACM Conference on Security and Privacy in Wireless and Mobile Networks, WiSec 2024
Country/Territory	Korea, Republic of
City	Seoul
Period	5/27/24 → 5/29/24

Keywords

human activity recognition
label poisoning
millimeter-wave (mmwave) technology
supervised contrastive learning (scl)

ASJC Scopus subject areas

Computer Networks and Communications
Computer Science Applications
Information Systems
Software
Safety Research

Access to Document

10.1145/3643833.3656123

Cite this

Singha, A., Bi, Z., Li, T., Chen, Y., & Zhang, Y. (2024). Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping. In WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks (pp. 31-41). (WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks). Association for Computing Machinery, Inc. https://doi.org/10.1145/3643833.3656123

Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping. / Singha, Amit; Bi, Ziqian; Li, Tao et al.
WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks. Association for Computing Machinery, Inc, 2024. p. 31-41 (WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks).

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

Singha, A, Bi, Z, Li, T, Chen, Y & Zhang, Y 2024, Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping. in WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks. WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks, Association for Computing Machinery, Inc, pp. 31-41, 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks, WiSec 2024, Seoul, Korea, Republic of, 5/27/24. https://doi.org/10.1145/3643833.3656123

Singha A, Bi Z, Li T, Chen Y, Zhang Y. Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping. In WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks. Association for Computing Machinery, Inc. 2024. p. 31-41. (WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks). doi: 10.1145/3643833.3656123

Singha, Amit ; Bi, Ziqian ; Li, Tao et al. / Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping. WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks. Association for Computing Machinery, Inc, 2024. pp. 31-41 (WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks).

@inproceedings{a4b0cc6ecc954f6eb273e87de1ce619c,

title = "Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping",

abstract = "Wireless Human Activity Recognition (HAR), leveraging their non-intrusive nature, has the potential to revolutionize various sectors, including healthcare, virtual reality, and surveillance. The advent of millimeter wave (mmWave) technology has significantly enhanced the capabilities of wireless HAR systems. This paper presents the first systematic study on the vulnerabilities of mmWave-based HAR to label flipping poisoning attacks in the context of supervised contrastive learning. We identify three label poisoning attacks on the contrastive mmWave-based HAR and propose corresponding countermeasures. The efficacy of the attacks and also our countermeasures are experimentally validated on a prototype system. The attacks and countermeasures can be easily extended to other wireless HAR systems, thereby promoting security considerations in system design and deployment.",

keywords = "human activity recognition, label poisoning, millimeter-wave (mmwave) technology, supervised contrastive learning (scl)",

author = "Amit Singha and Ziqian Bi and Tao Li and Yimin Chen and Yanchao Zhang",

note = "Publisher Copyright: {\textcopyright} 2024 Owner/Author.; 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks, WiSec 2024 ; Conference date: 27-05-2024 Through 29-05-2024",

year = "2024",

month = may,

day = "27",

doi = "10.1145/3643833.3656123",

language = "English (US)",

series = "WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks",

publisher = "Association for Computing Machinery, Inc",

pages = "31--41",

booktitle = "WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks",

}

TY - GEN

T1 - Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping

AU - Singha, Amit

AU - Bi, Ziqian

AU - Li, Tao

AU - Chen, Yimin

AU - Zhang, Yanchao

PY - 2024/5/27

Y1 - 2024/5/27

N2 - Wireless Human Activity Recognition (HAR), leveraging their non-intrusive nature, has the potential to revolutionize various sectors, including healthcare, virtual reality, and surveillance. The advent of millimeter wave (mmWave) technology has significantly enhanced the capabilities of wireless HAR systems. This paper presents the first systematic study on the vulnerabilities of mmWave-based HAR to label flipping poisoning attacks in the context of supervised contrastive learning. We identify three label poisoning attacks on the contrastive mmWave-based HAR and propose corresponding countermeasures. The efficacy of the attacks and also our countermeasures are experimentally validated on a prototype system. The attacks and countermeasures can be easily extended to other wireless HAR systems, thereby promoting security considerations in system design and deployment.

AB - Wireless Human Activity Recognition (HAR), leveraging their non-intrusive nature, has the potential to revolutionize various sectors, including healthcare, virtual reality, and surveillance. The advent of millimeter wave (mmWave) technology has significantly enhanced the capabilities of wireless HAR systems. This paper presents the first systematic study on the vulnerabilities of mmWave-based HAR to label flipping poisoning attacks in the context of supervised contrastive learning. We identify three label poisoning attacks on the contrastive mmWave-based HAR and propose corresponding countermeasures. The efficacy of the attacks and also our countermeasures are experimentally validated on a prototype system. The attacks and countermeasures can be easily extended to other wireless HAR systems, thereby promoting security considerations in system design and deployment.

KW - human activity recognition

KW - label poisoning

KW - millimeter-wave (mmwave) technology

KW - supervised contrastive learning (scl)

UR - http://www.scopus.com/inward/record.url?scp=85198107164&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85198107164&partnerID=8YFLogxK

U2 - 10.1145/3643833.3656123

DO - 10.1145/3643833.3656123

M3 - Conference contribution

AN - SCOPUS:85198107164

T3 - WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks

SP - 31

EP - 41

BT - WiSec 2024 - Proceedings of the 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks

PB - Association for Computing Machinery, Inc

T2 - 17th ACM Conference on Security and Privacy in Wireless and Mobile Networks, WiSec 2024

Y2 - 27 May 2024 through 29 May 2024

ER -

Securing Contrastive mmWave-based Human Activity Recognition against Adversarial Label Flipping

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this