ePrints@IIScePrints@IISc Home | About | Browse | Latest Additions | Advanced Search | Contact | Help

Plug-and-pipeline: Efficient regularization for single-step adversarial training

Vivek, BS and Revanur, A and Venkat, N and Babu, RV (2020) Plug-and-pipeline: Efficient regularization for single-step adversarial training. In: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2020, 14-19 June 2020, United States, pp. 138-146.

[img] PDF
iee_con_com_vis_pat_rec_wor_2020_138-146_2020.pdf - Published Version
Restricted to Registered users only

Download (464kB) | Request a copy
Official URL: https://dx.doi.org/10.1109/CVPRW50498.2020.00023


Adversarial Training (AT) is a straight forward solution to learn robust models by augmenting the training mini-batches with adversarial samples. Adversarial attack methods range from simple non-iterative (single-step) methods to computationally complex iterative (multi-step) methods. Although the single-step methods are efficient, the models trained using these methods merely appear to be robust, due to the masked gradients. In this work, we propose a novel regularizer named Plug-And-Pipeline (PAP) for single-step AT. The proposed regularizer attenuates the gradient masking effect by promoting the model to learn similar representations for both single-step and multi-step adversaries. Further, we present a novel pipelined approach that allows an efficient implementation of the proposed regularizer. Plug-And-Pipeline yields robustness comparable to multi-step AT methods, while requiring a low computational overhead, similar to that of single-step AT methods. © 2020 IEEE.

Item Type: Conference Paper
Publication: IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops
Publisher: IEEE Computer Society
Additional Information: The copyright of this article belongs to IEEE Computer Society
Keywords: Computer vision; Iterative methods, Attack methods; Computational overheads; Efficient implementation; Non-iterative; Regularizer; Robust models; Single-step; Single-step method, Pipelines
Department/Centre: Division of Interdisciplinary Sciences > Computational and Data Sciences
Date Deposited: 25 Sep 2020 10:59
Last Modified: 25 Sep 2020 10:59
URI: http://eprints.iisc.ac.in/id/eprint/66540

Actions (login required)

View Item View Item