Artificial Intelligence for improving cybersecurity efficiency

Point of contact

  • Benjamin Smith, INRIA-Saclay
  • Maryline Laurent, SAMOVAR-Telecom SudParis

Summary

Cybersecurity is a very dynamic sector, where attacks targeting systems and related countermeasures are continuously emerging. In this context, Artificial Intelligence, including Machine Learning or Deep Learning, can help to enhance the efficiency of cybersecurity countermeasures. Research investigations are still needed for assessing and enhancing cybersecurity measures: for generating large evaluation datasets and attacks, for detecting new botnets and malwares, for predicting new attacks, for deep packet inspection analysis, and for countermeasure decision making. Another direction known as adversarial AI analyses how an attacker provided with some AI capacities is able to repeatedly interact with a normal AI-based service to make it become unstable, and thus break the underlying AI model’s performance – resulting in inaccurate predictions.

Keywords

cybersecurity, artificial intelligence, adversarial AI

Researchers involved or interested

  • Gregory Blanc, Samovar
  • Christophe Kiennert, Samovar

Related master programs