C&ESAR 2025: Hugo Bourreau presented his work on digital twins

December 19, 2025December 19, 2025 By Dylan Massieux ECW / Event / PhD PresentationC&ESAR,cyber,cybercni,cybersecurity,ECW,PhD,research

The Cyber CNI Chair looks back at C&ESAR 2025, the cybersecurity conference organized by the Direction Générale de l’Armement (DGA) as part of European Cyber Week in Rennes. Among the notable presentations at this year’s event, one by Hugo Bourreau, a doctoral student jointly supervised by IMT Atlantique (France) and the University of Quebec at Chicoutimi (UQAC) (Canada), attracted particular attention.

On Wednesday, November 19, at 2:30 p.m., Hugo presented his work under the title:
“Machine Learning within Digital Twins for Threat Estimation and Detection (SoK),” highlighting the importance of his research for the cybersecurity of industrial and critical systems.

Focus on Digital Twins for cybersecurity

Hugo Bourreau‘s research lies at the intersection of artificial intelligence and cyber-physical systems (CPS). He has shown how Digital Twins—virtual replicas of real systems—can improve the detection, estimation, and prevention of cyberattacks.

Supervised by Fabien Dagnat, Marc-Oliver Pahl, Fehmi Jaafar, and Kevin Bouchard, Hugo combined physical modeling to understand the actual behavior of the system, behavioral analysis to detect anomalies, and secure machine learning to apply reliable and traceable algorithms.

Innovation areas presented

Hybrid detection : combining physical models and AI algorithms to identify attacks earlier and reduce false positives.
Enhanced monitoring : integrating physical parameters (temperature, vibrations, CPU load) with network signals to locate anomalies.
Digital substitution : replay attacks, train operators, and perform forensic analysis securely via the Digital Twin.
Reliable learning : separation of offline training on validated data and online inference, ensuring robustness and traceability.

Towards predictive Digital Twins

Hugo also presented the evolution of Digital Twins towards predictive models capable of anticipating malicious behavior before it occurs, paving the way for a higher level of resilience for critical industrial environments.

Recognition from the cyber community

Hugo’s selection at C&ESAR was a testament to the scientific quality and operational relevance of his work, illustrating the importance of linking academic research to the concrete challenges of cybersecurity.

Find out more :
🔗 https://cybercni.fr
🔗 https://www.european-cyber-week.eu
🔗 https://2025.cesar-conference.org