Brain programming is immune to adversarial attacks: Towards accurate and robust image classification using symbolic learning
Created by W.Langdon from
gp-bibliography.bib Revision:1.7975
- @Article{IBARRAVAZQUEZ:2022:swevo,
-
author = "Gerardo Ibarra-Vazquez and Gustavo Olague and
Mariana Chan-Ley and Cesar Puente and
Carlos Soubervielle-Montalvo",
-
title = "Brain programming is immune to adversarial attacks:
Towards accurate and robust image classification using
symbolic learning",
-
journal = "Swarm and Evolutionary Computation",
-
volume = "71",
-
pages = "101059",
-
year = "2022",
-
ISSN = "2210-6502",
-
DOI = "doi:10.1016/j.swevo.2022.101059",
-
URL = "https://www.sciencedirect.com/science/article/pii/S2210650222000311",
-
keywords = "genetic algorithms, genetic programming, Brain
programming, Adversarial attacks, Image classification,
Art media categorization",
-
abstract = "In recent years, the security concerns about the
vulnerability of deep convolutional neural networks to
adversarial attacks in slight modifications to the
input image almost invisible to human vision make their
predictions untrustworthy. Therefore, it is necessary
to provide robustness to adversarial examples with an
accurate score when developing a new classifier. In
this work, we perform a comparative study of the
effects of these attacks on the complex problem of art
media categorization, which involves a sophisticated
analysis of features to classify a fine collection of
artworks. We tested a prevailing bag of visual words
approach from computer vision, four deep convolutional
neural networks (AlexNet, VGG, ResNet, ResNet101), and
brain programming. The results showed that brain
programming predictions' change in accuracy was below
2percent using adversarial examples from the fast
gradient sign method. With a multiple-pixel attack,
brain programming obtained four out of seven classes
without changes and the rest with a maximum error of
4percent. Finally, brain programming got four
categories without changes using adversarial patches
and for the remaining three classes with an accuracy
variation of 1percent. The statistical analysis
confirmed that brain programming predictions'
confidence was not significantly different for each
pair of clean and adversarial examples in every
experiment. These results prove brain programming's
robustness against adversarial examples compared to
deep convolutional neural networks and the computer
vision method for the art media categorization
problem",
- }
Genetic Programming entries for
Gerardo Ibarra-Vazquez
Gustavo Olague
Mariana Chan-Ley
Cesar Puente
Carlos Soubervielle-Montalvo
Citations