Teaching and Supervision
Teaching
Since 2022, I take part in the coordination and conception of te SaclAI-School — BrevetAI project: an online platform to teach artificial intelligence at all levels. The goal of the platform is to make students understand AI concepts through interactive activities as much as possible, in contrast to standard text-and-video-based MOOCs. The platform, funded by the DataIA Institute, will be made available for all Université Paris-Saclay curricula starting from the 2024 — 2025 academic year.
Students
Current PhD students:
- Hovhannes Margaryan (2024–), Enabling subtle visual modifications with AR+AI (with Christian Sandor)
- Benjamin Loison (2024–), Image murmurs: forensic analysis of image noise patterns and structures to detect fake images, with Pablo Musé, Rafael Grompone, and Marina Gardella
- Matthieu Serfaty (2022–), Fighting misinformation in satellite imagery (with Rafael Grompone, Carlo de Franchis, Tina Nikoukhah)
- Yanhao Li (2021–), Deepfake detection by in-depth video structure analysis (with Jean-Michel Morel, Rafael Grompone, Miguel Colom, Tina Nikoukhah, Marina Gardella)
Master’s students:
- Benjamin Loison: Post-processing-robust PRNU analysis for reliable source camera identification in the smartphone era. on postprocessing-robust PRNU analysis for reliable source camera identification in the smartphone era, with Rafael Grompone, Pablo Musé, Marina Gardella, and in collaboration with Denis Perraud, Service National de Police Scientifique (SNPS, French scientific police)
- David Iagaru: Detecting AI-modified images with localized double compression detection, with Rafael Grompone and Yanhao Li
