会议论文

Luo, Wenjian

英文

2021 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC 2021)

2021

1792-1799

IEEE Congress on Evolutionary Computation (IEEE CEC)

IEEE Congress on Evolutionary Computation

JUN 28-JUL 01, 2021

ELECTR NETWORK

[Wu, Chenwang;Zhou, Nan;Xu, Peilan] Univ Sci & Technol China, Sch Comp Sci & Technol, Hefei 230027, Anhui, Peoples R China.^[Luo, Wenjian] Harbin Inst Technol, Sch Comp Sci & Technol, Shenzhen 518000, Guangdong, Peoples R China.^[Zhu, Tao] Univ South China, Sch Software Engn, Hengyang 421001, Hunan, Peoples R China.

IEEE, IEEE Computat Intelligence Soc, Jagiellonian Univ, Polish Acad Sci, Comm Informat, AGH Univ Sci & Technol, Warsaw Univ Technol

345 E 47TH ST, NEW YORK, NY 10017 USA

IEEE

978-1-7281-8392-3

10.1109/CEC45853.2021.9504790

National Key Research and Development Program of China [2020YFB2104003]; National Natural Science Foundation of ChinaNational Natural Science Foundation of China (NSFC) [61573327]

本校为其他机构

Studies have shown that deep neural networks (DNNs) are susceptible to adversarial attacks, which can cause misclassification. The adversarial attack problem can be regarded as an optimization problem, then the genetic algorithm (GA) that is problem-independent can naturally be designed to solve the optimization problem to generate effective adversarial examples. Considering the dimensionality curse in the image processing field, traditional genetic algorithms in high-dimensional problems often fall into local optima. Therefore, we propose a GA with multiple fitness functions (MF-GA). Specific...