Doctorant, Polytech Annecy-Chambéry.


Courriel :  michael.dell-aiera@univ-smb.fr

Téléphone : +33 (0)4 50 09 xx xx

Portable :

Bureau : A221

Adresse : LISTIC, 5 chemin de bellevue, CS80439, 74 944 Annecy Cedex



Theme : AFuTé

Sujet : Deep learning applied to Cherenkov Telescope Array images

Résumé :

Gamma-ray astronomy is a field of physics that studies astrophysics sources of high-energy photons. These uncharged particles travel in straight lines and are not deviated by magnetic fields along their journey to Earth, making possible the track of their origins. Once they interact with the atmosphere, they produce particle showers, emitting lights through the so-called Cherenkov process, which will be captured by the optical telescopes called Imaging Atmospheric Cherenkov Telescopes (IACT). Their analysis consists in solving an inverse problem : the determination of the energy and direction of arrival of the incoming photons with the telescope images. Through these studies, scientists may understand for instance the star birth process, the nature of dark matter, or even the evolution of galaxies. But, this is a complicated task because other cosmic particles produce the same kind of showers, and photons are highly under-represented. Fortunately, deep learning has been a great tool in solving computer vision tasks. However, the training procedure in our case implies the need of labeled data, which is unobtainable. Therefore, Monte Carlo simulations are used for the training, but the performance are degraded when applied on real data. This is the consequence of domain shifts between the two distributions. In this thesis, we will study how domain adaptation technics can be applied to this specific problem to improve the learning transfert from simulated data to observed data.

Mots-clefs : Gamma-ray astronomy, CTA, Deep learning, Domain adaptation.

Publications :(lien sous HAL)

Encadrant : Alexandre Benoit (LISTIC), Gilles Maurin (LAPP)

Co-encadrant : Thomas Vuillaume (LAPP), Flavien Vernier (LISTIC)

Début de la thèse : 01/10/2021

Ecole doctorale : SIE