EUR Cyberschool: Reflecting on the Benefits of Federated Learning for Large-Scale Network Incident Detection
On April 14, 2023, Léo Lavaur and Yann Busnel, presented their work titled “The Value of Federated Learning in the Context of Incident Detection on Large-Scale Networks and/or Systems” at the Spring School of EUR Cyberschool .
The Spring School of EUR Cyberschool took place during mid of April. The school is dedicated to providing an overview of the current state of research in cybersecurity. The event featured presentations from recent works by researchers from IRISA, IETR, IRMAR, and CREST. This first edition was open to local cybersecurity PhD students, as well as students from the University of Rennes and partner engineering schools of the CyberSchool.
During the presentation, Léo and Yann discussed the history and benefits of federated learning for securing large-scale networks and/or collaborating among multiple distinct networks. Léo then hosted a hands-on workshop on federated learning for intrusion detection. Participants had the opportunity to experiment with federated learning on a provided dataset and compare it with traditional machine learning approaches.
Léo Lavaur, a PhD student at the Cyber CNI Chair at IMT Atlantique and a member of the SOTERN team at IRISA, is a specialist in federated learning and anomaly detection for critical infrastructure security. Yann Busnel, a professor at IMT Atlantique and a researcher in the SOTERN team at IRISA, is an expert in distributed algorithms and large-scale system modeling.
We hope that those who attended found the insights on cybersecurity and federated learning from experts in the field valuable. Stay tuned for future events and presentations on this important subject!
To learn more:
- The notebook of the workshop can be found here.
About the Spring School of EUR Cyberschool
The Spring School of EUR Cyberschool took place during the last week. The school is dedicated to providing an overview of the current state of research in cybersecurity. The event featured presentations from recent works by researchers from IRISA, IETR, IRMAR, and CREST. This first edition was open to local cybersecurity PhD students, as well as students from the University of Rennes and partner engineering schools of the CyberSchool.
I have obtained my PhD from the University of Strasbourg, where I worked in the Networks Team of the ICube laboratory. During this time,
I worked mainly on data leak prevention in multi party workflows, but also on related aspects like access control, and on using metagraphs to represent security policies. In addition, I have worked on complexity proofs with hypergraphs.
Before my PhD, I obtained my master’s degree in Computer Networks and Embedded Systems from the University of Strasbourg. In the finalyear internship, I worked on network security in the Internet, and more specifically on the security of the BGP protocol. Those experiences have allowed me to gather a deep understanding of topics related to networking and cybersecurity in general. I have also been involved in the research community in general, giving courses to students, being part of the article review process, and being part of the organizing committee for conferences.
- EUR Cyberschool: Reflecting on the Benefits of Federated Learning for Large-Scale Network Incident Detection - June 8, 2023
- “Blockchain for Collaborative Cybersecurity” – Our PostDoc Loïc Miller talks at the EUR Cyberschool - June 7, 2023
- Salim-Amine BOU ARAM is becoming a reviewer for IETF Interne Area Directorate group - April 10, 2023
