2026.05.18Paper presented at IEEE/IFIP NOMS 2026 (MCT workshop) on a graph-based approach (GPML) for detecting cyberattacks in water distribution networks from topology changes.
Julien Michel defended his PhD thesis, 'Real-time robust attack detection using graph community metrics', on 8 April 2026 at the University of Strasbourg.
Majed Jaber defended his PhD thesis, 'Structural and spectral analysis of dynamic graphs for attack detection', on 19 December 2025 at the University of Strasbourg.
Paper presented at KES 2025 (Osaka) defining t-robustness and a feature engineering approach, evaluated on UGR16 enriched with graph community metrics, that produces attack-detection models more stable under concept drift.
2025.09.01Journal article (SoftwareX) presenting GPML, an open-source library that transforms raw network traffic into graph representations for community and spectral-metric-based cyberattack detection.
GPML has for purpose to give an easy to use tool to ready data for use to machine learning algorithm using graph modelisation to compute new features and insert them to pandas dataframe. It also comes with a module with one ready function to use to convert dataset to html graph representation.
2023.01.01XDGMed develops explainable graph-based methods for detecting attacks in medical and other sensitive dynamic systems, combining graph analysis with trusted artificial intelligence principles.
DAMIAGE develops active, graph-based attack detection for large-scale critical infrastructures, combining network traffic with operational signals to support real-time or near-real-time cybersecurity.