Bio

Me in brief:

• Deputy head of LRE lab at EPITA since 9/2022
• Head of Security and Systems group of the LRE ([was previously LSE] (https://www.lse.epita.fr/) until 9/2022) since 4/2022
• Member of [CSTB - Complex Systems and Translational Bioinformatics - team] (https://cstb.icube.unistra.fr/) at [ICube Laboratory, University of Strasbourg] (https://icube.unistra.fr/) since 2012
• Full Professor at EPITA Strasbourg since 5/2021

Research Area

My research work is drawn by one core question:

How to detect complex attacks in dynamic digital environments that generate huge data volumes

And focuses on 3 challenges:

• How to detect complex, multi-step attacks in system traces ?
• How to learn novel attacks to adapt analysis and prepare reaction ?
• How to model these attacks for an explainable, transferable detection ?

Traditional attack models are inherently static.

Machine learning approaches bring classification and anomaly detection capabilities, but focus on individual data points (i.e. packets) have huge limitation in considering the interactions between machines, services, users.

Since my arrival at EPITA I focus on the study of graph models as explainable vectors for detection of attacks in heterogeneous environments.

Graph models provide a powerfull approach to support:

• Interaction between machines
• Analysis of attack paths and scenarios, PhD Thesis of Julio Navarro Lara, defended on 2019 at University of Strasbourg
• Explainability - ANR-IA THIA-Artic Project, PhD thesis of Amani Abou Rida
• Scalability - DGA RAPID Damiage Project, PhD thesis of Julien Michel
• Dynamicity - Contrat Doctoral Région Grand-Est / EPITA, XDGMed Project, PhD Thesis of Majed Jaber

The analysis of complex attacks require stocastic approaches. This is the cas of graph learning, as it is the case of evolutionary computation, which I also explore like in ANR Correau - Résilience par la conception et sécurité des réseaux d’eau - Project.