Home > Research > Publications & Outputs > GradMDM

Electronic data

  • Layer_wise_Attack

    Rights statement: ©2023 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

    Accepted author manuscript, 861 KB, PDF document

    Available under license: CC BY: Creative Commons Attribution 4.0 International License

Links

Text available via DOI:

View graph of relations

GradMDM: Adversarial Attack on Dynamic Networks

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

GradMDM: Adversarial Attack on Dynamic Networks. / Pan, Jianhong; Foo, Lin Geng ; Zheng, Qichen et al.
In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 45, No. 9, 01.09.2023, p. 11374-11381.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Pan, J, Foo, LG, Zheng, Q, Fan, Z, Rahmani, H, Ke, Q & Liu, J 2023, 'GradMDM: Adversarial Attack on Dynamic Networks', IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 9, pp. 11374-11381. https://doi.org/10.1109/TPAMI.2023.3263619

APA

Pan, J., Foo, L. G., Zheng, Q., Fan, Z., Rahmani, H., Ke, Q., & Liu, J. (2023). GradMDM: Adversarial Attack on Dynamic Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(9), 11374-11381. https://doi.org/10.1109/TPAMI.2023.3263619

Vancouver

Pan J, Foo LG, Zheng Q, Fan Z, Rahmani H, Ke Q et al. GradMDM: Adversarial Attack on Dynamic Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2023 Sept 1;45(9):11374-11381. Epub 2023 Mar 31. doi: 10.1109/TPAMI.2023.3263619

Author

Pan, Jianhong ; Foo, Lin Geng ; Zheng, Qichen et al. / GradMDM : Adversarial Attack on Dynamic Networks. In: IEEE Transactions on Pattern Analysis and Machine Intelligence. 2023 ; Vol. 45, No. 9. pp. 11374-11381.

Bibtex

@article{db017c87f2a040f59364e6bce2ef2064,
title = "GradMDM: Adversarial Attack on Dynamic Networks",
abstract = "Dynamic neural networks can greatly reduce computation redundancy without compromising accuracy by adapting their structures based on the input. In this paper, we explore the robustness of dynamic neural networks against \textit{energy-oriented attacks} targeted at reducing their efficiency.Specifically, we attack dynamic models with our novel algorithm GradMDM.GradMDM is a technique that adjusts the direction and the magnitude of the gradients to effectively find a small perturbation for each input, that will activate more computational units of dynamic models during inference. We evaluate GradMDM on multiple datasets and dynamic models, where it outperforms previous energy-oriented attack techniques, significantly increasing computation complexity while reducing the perceptibility of the perturbations. ",
author = "Jianhong Pan and Foo, {Lin Geng} and Qichen Zheng and Zhipeng Fan and Hossein Rahmani and Qiuhong Ke and Jun Liu",
note = "{\textcopyright}2023 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE. ",
year = "2023",
month = sep,
day = "1",
doi = "10.1109/TPAMI.2023.3263619",
language = "English",
volume = "45",
pages = "11374--11381",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "9",

}

RIS

TY - JOUR

T1 - GradMDM

T2 - Adversarial Attack on Dynamic Networks

AU - Pan, Jianhong

AU - Foo, Lin Geng

AU - Zheng, Qichen

AU - Fan, Zhipeng

AU - Rahmani, Hossein

AU - Ke, Qiuhong

AU - Liu, Jun

N1 - ©2023 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

PY - 2023/9/1

Y1 - 2023/9/1

N2 - Dynamic neural networks can greatly reduce computation redundancy without compromising accuracy by adapting their structures based on the input. In this paper, we explore the robustness of dynamic neural networks against \textit{energy-oriented attacks} targeted at reducing their efficiency.Specifically, we attack dynamic models with our novel algorithm GradMDM.GradMDM is a technique that adjusts the direction and the magnitude of the gradients to effectively find a small perturbation for each input, that will activate more computational units of dynamic models during inference. We evaluate GradMDM on multiple datasets and dynamic models, where it outperforms previous energy-oriented attack techniques, significantly increasing computation complexity while reducing the perceptibility of the perturbations.

AB - Dynamic neural networks can greatly reduce computation redundancy without compromising accuracy by adapting their structures based on the input. In this paper, we explore the robustness of dynamic neural networks against \textit{energy-oriented attacks} targeted at reducing their efficiency.Specifically, we attack dynamic models with our novel algorithm GradMDM.GradMDM is a technique that adjusts the direction and the magnitude of the gradients to effectively find a small perturbation for each input, that will activate more computational units of dynamic models during inference. We evaluate GradMDM on multiple datasets and dynamic models, where it outperforms previous energy-oriented attack techniques, significantly increasing computation complexity while reducing the perceptibility of the perturbations.

U2 - 10.1109/TPAMI.2023.3263619

DO - 10.1109/TPAMI.2023.3263619

M3 - Journal article

VL - 45

SP - 11374

EP - 11381

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 9

ER -