Research

Research

My main interest is in using advances in machine learning for the measure of pollution and biodiversity. I At Centre de Mathématiques Appliquées (CMAP) - Institut Polytechnique de Paris (IPP), I worked on hybrid deep learning: combining models (statistical, physical, etc.) with modern deep learning architectures.

Recent interests:

  • Medical Image analysis & denoising (project with Jean Ollion)
  • Object tracking via modelisation of Camera tracking
  • Including MCMC and SMC in Deep Learning models such as VAEs, Transformers…

Teaching

I have taught Deep Learning in various institutions (IPP, Centrale-Supelec, EPITA…) since 2015. All my teaching material (lectures, code, notebooks, videos) is available freely on this page.

Students

  • Hedi ben Younes, LIP6, PhD co-advisor, 2015-2019, “Apprentissage de représentations multimodales texte-images”
  • Arthur Douillard, LIP6, PhD co-advisor, 2019-2022, “Apprentissage profond incremental avec peu de donnees et une grande variabilite inter-domaines”
  • Mathis Chagneux, AgroParisTech, PhD co-advisor, 2020-2023, “calibration de modèles géospatiaux robustes pour le comptage, le suivi et la localisation de déchets plastiques”

Publications

DLPB Scholar