Saut quantique : Exploration de la cybersécurité des réseaux électriques intelligents avec l'apprentissage quantique

We will first start by introducing smart grid cybersecurity challenges and explaining attacks on transmission line protection systems, showing how malicious actors can target substations through various attack scenarios. Next, the presentation will outline our project objectives, including developing classical ML baselines for attack detection and quantum ML equivalents. The core content will cover our methodology using Siamese Neural Networks for handling imbalanced data, followed by enhancements through Snake activation functions and AutoML techniques. Finally, we'll introduce quantum activation functions that exploit entanglement to detect correlations missed by classical methods, present comparative results showing quantum methods outperforming classical approaches on surrogate problems, and conclude with future directions for incorporating quantum techniques into AutoML search spaces.