GammaLearn

GammaLearn is a collaborative project to apply deep learning to the analysis of low-level Imaging Atmospheric Cherenkov Telescopes such as CTA. The project focus both on improving reconstruction performances but also providing an environment for the analysis of CTA data with DL.

Progress update

Continuation of the project started in 2017: investigation of deep learning applied to the Cherenkov Telescope Array.

• Efforts to understand better and improve the results obtained for single-telescope reconstruction
• Investigation of the influence of the cleaning on the reconstruction performances
• Production of reference curves with a standard analysis pipeline for single-telescope reconstruction
• Presentations and hands-on in several CTA meetings

Challenges at the beginning of the project

CTA specific images are not adapted to deep learning framework

Innovative solution produced

• Release of the Indexed Convolution library (https://indexed-convolution.readthedocs.io)
• Peer-reviewed publication of the indexed convolution method at VISAPP 2019 (Jacquemont et al, 2019)

Future activities

Continuation of the GammaLearn project under ESCAPE2020

CONTACT

Thomas Vuillaume