E3S Web Conf.
Volume 229, 2021The 3rd International Conference of Computer Science and Renewable Energies (ICCSRE’2020)
|Number of page(s)||9|
|Published online||25 January 2021|
Analysis of Security of Machine Learning and a proposition of assessment pattern to deal with adversarial attacks
Laboratory of Electrical and Telecommunication Engineering Ibn Tofail science university Kenitra, Morocco
Today, Machine Learning is being rolled out in a variety of areas. It is a promising field that can offer several assets and can revolutionize several aspects of technology. Nevertheless, despite the advantages of machine learning technologies, learning algorithms can be exploited by attackers to carry out illicit activities. Therefore, the field of security of machine learning is deriving attention in these times so as to meet this challenge and develop secure learning models. In this paper, we overview a taxonomy that will help us understand and analyze the security of machine learning models. In the next sections, we conduct a comparative study of most widespread adversarial attacks then, we analyze common methods that were advanced to protect systems built on Machine learning models from adversaries. Finally, we discuss a proposition of a pattern designed to ensure a security assessment of machine learning models.
Key words: Machine Learning / Adversarial Attacks / Mitigation techniques.
© The Authors, published by EDP Sciences, 2021
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.