Explainable Artificial Intelligence for Online Planning of Cyber-Attack Response

To counter cyber-attackers with access to ever more powerful and commoditized Artificial Intelligence (AI), next generation SOAR (Security Orchestration Automation and Response) frameworks assisting SOC (Security Operations Center) cybersecurity experts to detect and respond to cyberattacks will need to integrate a comprehensive set of explainable AI services. The thesis topic is the engineering of two such AI services, one for online planning of cyberattack response and one for online planning of cyberattacks, co-trained against one another through adversarial machine learning in red team vs. blue team attack-defense simulations. 

The starting hypothesis for the research is that both these explainable AI services can be rapidly implemented and tested, partially through declarative manual programming, and partially through adversarial machine learning, by probabilistic logic programs with multi-objective utility facts and probabilistic event calculus metarules.

The Engineering School of Informatics Electronics and Automation (ESIEA) https://www.esiea.fr/en/ offers undergraduate and graduate programs in computer science, computer engineering, information technology and cybersecurity on 5 French campuses in Paris, Ivry (a Paris suburb), Laval (in the northwest of France), Dax and Agen (both in the southwest of France). 

Founded in 1958, ESIEA is non-profit private higher education institution, governed by its alumni and recognized by the French government for its general usefullness to French society. 

The thesis will be conducted in ESIEA's Learning, Data and Robotics research team https://www.esiea.fr/pedagogie/laboratoires/ldr/and in close collaboration with the Center for Research in Informatics (CRI) of University Paris 1 Panthéon-Sorbonne https://cri.pantheonsorbonne.fr/

• Holder of a Master's or Engineering Degree in Computer Science or Operations Research. 
• Excellent computer programming skills.
• Either a strong background in symbolic AI (knowledge representation, logic and/or constraint programming, Bayesian reasoning) with the intellectual curiosity to acquire a focused expertise in cybersecurity (incident response, malware analysis, threat intelligence) during the thesis
• Or, vice-versa, a strong background in cybersecurity (incident response, malware analysis, threat intelligence) with the intellectual curiosity to acquire a focused expertise in symbolic AI (probabilistic logic programming) during the thesis