GreedyFool: Multi-factor imperceptibility and its application to designing a black-box adversarial attack

Liu, Hui; Zhao, Bo; Ji, Minzhi; Li, Mengchen; Liu, Peng

doi:10.1016/j.ins.2022.08.026

Cited by 5 publications

(2 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It firstly added a salt-and-pepper noise that could fool the model, then repeatedly removed the noise of one pixel if the model was still fooled. The GreedyFool method [120] developed a two-stage algorithm to minimize the 0 norm. The first stage increased the perturbed pixels according to the distortion map, and the second stage gradually reduced the perturbed pixels according to the attack performance with different perturbation magnitudes on these pixels.…”

Section: ) Comparison Between Optimization-based and Learning-mentioning

confidence: 99%

“…Xu et al [211] ICLR 2019 Su et al [172] IEEE TEC 2019 CD-UAP [228] AAAI 2020 Xu et al [212] ECCV 2020 Wu et al [202] ECCV 2020 SAPF [60] ECCV 2020 UPC [82] CVPR 2020 AdvCam [56] CVPR 2020 LG-GAN [246] CVPR 2020 Xu et al [213] CVPR 2020 PhysGAN [97] CVPR 2020 Li et al [106] CVPR 2020 GreedyFool [120] NeurIPS 2020 Xu et al [214] MM 2020 MAG-GAN [33] Information Sciences 2020 GUAP [234] ICDM 2020 Wong et al [197] ICLR 2021…”

Section: Tablementioning

confidence: 99%

See 1 more Smart Citation

Adversarial Machine Learning: A Systematic Survey of Backdoor Attack, Weight Attack and Adversarial Example

Liu¹,

Zhu²,

Liu³

et al. 2023

Preprint

View full text Add to dashboard Cite

Adversarial machine learning (AML) studies the adversarial phenomenon of machine learning, which may make inconsistent or unexpected predictions with humans. Some paradigms have been recently developed to explore this adversarial phenomenon occurring at different stages of a machine learning system, such as training-time adversarial attack (i.e., backdoor attack), deployment-time adversarial attack (i.e., weight attack), and inference-time adversarial attack (i.e., adversarial example). However, although these paradigms share a common goal, their developments are almost independent, and there is still no big picture of AML. In this work, we aim to provide a unified perspective to the AML community to systematically review the overall progress of this field. We firstly provide a general definition about AML, and then propose a unified mathematical framework to covering existing attack paradigms. According to the proposed unified framework, we can not only clearly figure out the connections and differences among these paradigms, but also systematically categorize and review existing works in each paradigm.

show abstract

Section: ) Comparison Between Optimization-based and Learning-mentioning

confidence: 99%