Our student Matisse Callewaert won the Leo Baekeland award 2025 for his master's thesis "Real-Time Adaptive Feature Extraction for ML-Based
Network Intrusion Detection".
In our cybersecurity research, we focus on the use of AI for network intrusion detection, specifically for detecting cyberattacks on corporate networks, where the attack may have never been seen before and thus cannot be detected by traditional signature-based systems. This research requires high-quality data on the network flows that run over a corporate network. For each of these, often encrypted, network flows, approximately 80 characteristics must be extracted (bitrate, burstiness, jitter, average packet size, etc.) in an extremely fast and reliable way, so that data traffic is not delayed. In his master's thesis, Matisse researched in detail how this can be achieved in soft real-time on high-speed network links and achieved excellent results.
