Abstract

This paper studies the security aspect of gossip-based decentralized optimization algorithms for multi agent systems against data injection attacks. Our contributions are two-fold. First, we show that the popular distributed projected gradient method (by Nedić et al.) can be attacked by coordinated insider attacks, in which the attackers are able to steer the final state to a point of their choosing. Second, we propose a metric that can be computed locally by the trustworthy agents processing their own iterates and those of their neighboring agents. This metric can be used by the trustworthy agents to detect and localize the attackers. We conclude the paper by supporting our findings with numerical experiments.

Original languageEnglish (US)
Title of host publication2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3644-3648
Number of pages5
Volume2018-April
ISBN (Print)9781538646588
DOIs
StatePublished - Sep 10 2018
Event2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Calgary, Canada
Duration: Apr 15 2018Apr 20 2018

Other

Other2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018
CountryCanada
CityCalgary
Period4/15/184/20/18

Fingerprint

Gradient methods
Multi agent systems
Processing
Experiments

Keywords

  • Data injection attack
  • Decentralized optimization
  • Gossip algorithms

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Xiaoxiao Wu, S., Wai, H. T., Scaglione, A., Nedich, A., & Leshem, A. (2018). Data injection attack on decentralized optimization. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings (Vol. 2018-April, pp. 3644-3648). [8462528] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2018.8462528

Data injection attack on decentralized optimization. / Xiaoxiao Wu, Sissi; Wai, Hoi To; Scaglione, Anna; Nedich, Angelia; Leshem, Amira.

2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. p. 3644-3648 8462528.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Xiaoxiao Wu, S, Wai, HT, Scaglione, A, Nedich, A & Leshem, A 2018, Data injection attack on decentralized optimization. in 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. vol. 2018-April, 8462528, Institute of Electrical and Electronics Engineers Inc., pp. 3644-3648, 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018, Calgary, Canada, 4/15/18. https://doi.org/10.1109/ICASSP.2018.8462528
Xiaoxiao Wu S, Wai HT, Scaglione A, Nedich A, Leshem A. Data injection attack on decentralized optimization. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April. Institute of Electrical and Electronics Engineers Inc. 2018. p. 3644-3648. 8462528 https://doi.org/10.1109/ICASSP.2018.8462528
Xiaoxiao Wu, Sissi ; Wai, Hoi To ; Scaglione, Anna ; Nedich, Angelia ; Leshem, Amira. / Data injection attack on decentralized optimization. 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. pp. 3644-3648
@inproceedings{7cf92d75cd4e44b480b929851ea7a12e,
title = "Data injection attack on decentralized optimization",
abstract = "This paper studies the security aspect of gossip-based decentralized optimization algorithms for multi agent systems against data injection attacks. Our contributions are two-fold. First, we show that the popular distributed projected gradient method (by Nedić et al.) can be attacked by coordinated insider attacks, in which the attackers are able to steer the final state to a point of their choosing. Second, we propose a metric that can be computed locally by the trustworthy agents processing their own iterates and those of their neighboring agents. This metric can be used by the trustworthy agents to detect and localize the attackers. We conclude the paper by supporting our findings with numerical experiments.",
keywords = "Data injection attack, Decentralized optimization, Gossip algorithms",
author = "{Xiaoxiao Wu}, Sissi and Wai, {Hoi To} and Anna Scaglione and Angelia Nedich and Amira Leshem",
year = "2018",
month = "9",
day = "10",
doi = "10.1109/ICASSP.2018.8462528",
language = "English (US)",
isbn = "9781538646588",
volume = "2018-April",
pages = "3644--3648",
booktitle = "2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Data injection attack on decentralized optimization

AU - Xiaoxiao Wu, Sissi

AU - Wai, Hoi To

AU - Scaglione, Anna

AU - Nedich, Angelia

AU - Leshem, Amira

PY - 2018/9/10

Y1 - 2018/9/10

N2 - This paper studies the security aspect of gossip-based decentralized optimization algorithms for multi agent systems against data injection attacks. Our contributions are two-fold. First, we show that the popular distributed projected gradient method (by Nedić et al.) can be attacked by coordinated insider attacks, in which the attackers are able to steer the final state to a point of their choosing. Second, we propose a metric that can be computed locally by the trustworthy agents processing their own iterates and those of their neighboring agents. This metric can be used by the trustworthy agents to detect and localize the attackers. We conclude the paper by supporting our findings with numerical experiments.

AB - This paper studies the security aspect of gossip-based decentralized optimization algorithms for multi agent systems against data injection attacks. Our contributions are two-fold. First, we show that the popular distributed projected gradient method (by Nedić et al.) can be attacked by coordinated insider attacks, in which the attackers are able to steer the final state to a point of their choosing. Second, we propose a metric that can be computed locally by the trustworthy agents processing their own iterates and those of their neighboring agents. This metric can be used by the trustworthy agents to detect and localize the attackers. We conclude the paper by supporting our findings with numerical experiments.

KW - Data injection attack

KW - Decentralized optimization

KW - Gossip algorithms

UR - http://www.scopus.com/inward/record.url?scp=85054284797&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054284797&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2018.8462528

DO - 10.1109/ICASSP.2018.8462528

M3 - Conference contribution

AN - SCOPUS:85054284797

SN - 9781538646588

VL - 2018-April

SP - 3644

EP - 3648

BT - 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -