Summary of the presentation: In recent years, the Industry of the Future has experienced a significant digitalization of its systems to gain competitiveness. The integration of new technologies in these environments such as Manufacturing Cloud and Industrial IoT, has greatly increased the attack surface of these systems. To detect cyber-attacks that target this type of environment, the use of Artificial Intelligence in Intrusion Detection Systems (IDS) has become a necessity. However, various challenges need to be addressed to effectively integrate a machine learning algorithm into an IDS and detect complex and persistent attacks. This presentation aims to present different strategies to optimize the different steps necessary to design a machine learning model in order to effectively detect weak signals such as cyber-attacks. A focus on the importance of minimizing false positives will also be carried out.

Biography: David Espes received the MS degree in Computer Sciences (2004) and the PhD degree in Computer Sciences (2008) all from the University of Toulouse, France. In September of 2009, he joined the “Network, Security and Multimedia” department at Telecom Bretagne (Rennes) as a post-doctoral researcher. From 2010 to 2022, he was an associate professor at the University of Brest, France. Since 2022, he has been with the University of Brest, France, as a full professor.  He is the leader of the CNRS LabSTICC IRIS team on Security and Resilience of Information Systems. He works on cybersecurity in the field of Industry of the Future and Unmanned Aerial Vehicles, topics including management and deployment of security policies, intrusion detection and response to cyber-attacks.