A Distributed Biased Boundary Attack Method in Black-Box Attack

The adversarial samples threaten the effectiveness of machine learning (ML) models and algorithms in many applications. In particular, black-box attack methods are quite close to actual scenarios. Research on black-box attack methods and the generation of adversarial samples is helpful to discover t...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Fengtao Xiang, Jiahui Xu, Wanpeng Zhang, Weidong Wang
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	adversarial samples black-box attacks machine learning models boundary attacks Technology T Engineering (General). Civil engineering (General) TA1-2040 Biology (General) QH301-705.5 Physics QC1-999 Chemistry QD1-999
Acceso en línea:	https://doaj.org/article/31fcb370a0354069ae9989ccb021af4e
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:31fcb370a0354069ae9989ccb021af4e
record_format	dspace
spelling	oai:doaj.org-article:31fcb370a0354069ae9989ccb021af4e2021-11-11T15:25:39ZA Distributed Biased Boundary Attack Method in Black-Box Attack10.3390/app1121104792076-3417https://doaj.org/article/31fcb370a0354069ae9989ccb021af4e2021-11-01T00:00:00Zhttps://www.mdpi.com/2076-3417/11/21/10479https://doaj.org/toc/2076-3417The adversarial samples threaten the effectiveness of machine learning (ML) models and algorithms in many applications. In particular, black-box attack methods are quite close to actual scenarios. Research on black-box attack methods and the generation of adversarial samples is helpful to discover the defects of machine learning models. It can strengthen the robustness of machine learning algorithms models. Such methods require queries frequently, which are less efficient. This paper has made improvements in the initial generation and the search for the most effective adversarial examples. Besides, it is found that some indicators can be used to detect attacks, which is a new foundation compared with our previous studies. Firstly, the paper proposed an algorithm to generate initial adversarial samples with a smaller L<sub>2</sub> norm; secondly, a combination between particle swarm optimization (PSO) and biased boundary adversarial attack (BBA) is proposed. It is the PSO-BBA. Experiments are conducted on the ImageNet. The PSO-BBA is compared with the baseline method. Experimental comparison results certificate that: (1) A distributed framework for adversarial attack methods is proposed; (2) The proposed initial point selection method can reduces query numbers effectively; (3) Compared to the original BBA, the proposed PSO-BBA algorithm accelerates the convergence speed and improves the accuracy of attack accuracy; (4) The improved PSO-BBA algorithm has preferable performance on targeted and non-targeted attacks; (5) The mean structural similarity (MSSIM) can be used as the indicators of adversarial attack.Fengtao XiangJiahui XuWanpeng ZhangWeidong WangMDPI AGarticleadversarial samplesblack-box attacksmachine learning modelsboundary attacksTechnologyTEngineering (General). Civil engineering (General)TA1-2040Biology (General)QH301-705.5PhysicsQC1-999ChemistryQD1-999ENApplied Sciences, Vol 11, Iss 10479, p 10479 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	adversarial samples black-box attacks machine learning models boundary attacks Technology T Engineering (General). Civil engineering (General) TA1-2040 Biology (General) QH301-705.5 Physics QC1-999 Chemistry QD1-999
spellingShingle	adversarial samples black-box attacks machine learning models boundary attacks Technology T Engineering (General). Civil engineering (General) TA1-2040 Biology (General) QH301-705.5 Physics QC1-999 Chemistry QD1-999 Fengtao Xiang Jiahui Xu Wanpeng Zhang Weidong Wang A Distributed Biased Boundary Attack Method in Black-Box Attack
description	The adversarial samples threaten the effectiveness of machine learning (ML) models and algorithms in many applications. In particular, black-box attack methods are quite close to actual scenarios. Research on black-box attack methods and the generation of adversarial samples is helpful to discover the defects of machine learning models. It can strengthen the robustness of machine learning algorithms models. Such methods require queries frequently, which are less efficient. This paper has made improvements in the initial generation and the search for the most effective adversarial examples. Besides, it is found that some indicators can be used to detect attacks, which is a new foundation compared with our previous studies. Firstly, the paper proposed an algorithm to generate initial adversarial samples with a smaller L<sub>2</sub> norm; secondly, a combination between particle swarm optimization (PSO) and biased boundary adversarial attack (BBA) is proposed. It is the PSO-BBA. Experiments are conducted on the ImageNet. The PSO-BBA is compared with the baseline method. Experimental comparison results certificate that: (1) A distributed framework for adversarial attack methods is proposed; (2) The proposed initial point selection method can reduces query numbers effectively; (3) Compared to the original BBA, the proposed PSO-BBA algorithm accelerates the convergence speed and improves the accuracy of attack accuracy; (4) The improved PSO-BBA algorithm has preferable performance on targeted and non-targeted attacks; (5) The mean structural similarity (MSSIM) can be used as the indicators of adversarial attack.
format	article
author	Fengtao Xiang Jiahui Xu Wanpeng Zhang Weidong Wang
author_facet	Fengtao Xiang Jiahui Xu Wanpeng Zhang Weidong Wang
author_sort	Fengtao Xiang
title	A Distributed Biased Boundary Attack Method in Black-Box Attack
title_short	A Distributed Biased Boundary Attack Method in Black-Box Attack
title_full	A Distributed Biased Boundary Attack Method in Black-Box Attack
title_fullStr	A Distributed Biased Boundary Attack Method in Black-Box Attack
title_full_unstemmed	A Distributed Biased Boundary Attack Method in Black-Box Attack
title_sort	distributed biased boundary attack method in black-box attack
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/31fcb370a0354069ae9989ccb021af4e
work_keys_str_mv	AT fengtaoxiang adistributedbiasedboundaryattackmethodinblackboxattack AT jiahuixu adistributedbiasedboundaryattackmethodinblackboxattack AT wanpengzhang adistributedbiasedboundaryattackmethodinblackboxattack AT weidongwang adistributedbiasedboundaryattackmethodinblackboxattack AT fengtaoxiang distributedbiasedboundaryattackmethodinblackboxattack AT jiahuixu distributedbiasedboundaryattackmethodinblackboxattack AT wanpengzhang distributedbiasedboundaryattackmethodinblackboxattack AT weidongwang distributedbiasedboundaryattackmethodinblackboxattack
_version_	1718435299178053632

A Distributed Biased Boundary Attack Method in Black-Box Attack

Ejemplares similares