Towards a Certification of Deep Image Classifiers against Convolutional Attacks

Mallek Mziou-Sallami, Mallek Mziou-Sallami, Faouzi Adjed, Faouzi Adjed

2022

Abstract

Deep learning models do not achieve sufficient confidence, explainability and transparency levels to be integrated into safety-critical systems. In the context of DNN-based image classifier, robustness have been first studied under simple image attacks (2D rotation, brightness), and then, subsequently, under other geometrical perturbations. In this paper, we intend to introduce a new method to certify deep image classifiers against convolutional attacks. Using the abstract interpretation theory, we formulate the lower and upper bounds with abstract intervals to support other classes of advanced attacks including image filtering. We experiment the proposed method on MNIST and CIFAR10 databases and several DNN architectures. The obtained results show that convolutional neural networks are more robust against filtering attacks. Multilayered perceptron robustness decreases when increasing number of neurons and hidden layers. These results prove that the complexity of DNN models improves prediction’s accuracy but often impacts robustness.

Paper Citation

in Harvard Style

Mziou-Sallami M. and Adjed F. (2022). Towards a Certification of Deep Image Classifiers against Convolutional Attacks. In Proceedings of the 14th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART, ISBN 978-989-758-547-0, pages 419-428. DOI: 10.5220/0010870400003116

in Bibtex Style

@conference{icaart22,
author={Mallek Mziou-Sallami and Faouzi Adjed},
title={Towards a Certification of Deep Image Classifiers against Convolutional Attacks},
booktitle={Proceedings of the 14th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,},
year={2022},
pages={419-428},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010870400003116},
isbn={978-989-758-547-0},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,
TI - Towards a Certification of Deep Image Classifiers against Convolutional Attacks
SN - 978-989-758-547-0
AU - Mziou-Sallami M.
AU - Adjed F.
PY - 2022
SP - 419
EP - 428
DO - 10.5220/0010870400003116