Speaker
Description
CSES (China Seismo-Electromagnetic Satellite) is a sophisticated multi-payload space observatory aimed at the observation of the Van Allen Belts dynamics, the study of solar-terrestrial interactions and the extension at low energies of existing Cosmic Ray measurements. HEPD-02 is the new generation of High Energy Particle Detector and will be launched on board the China Seismo-Electromagnetic Satellite (CSES-02) at the end of this year. The instrument is optimized to detect fluxes of charged particles: mostly electrons and protons, with good capabilities in the identification of heavier nuclei and will provide measurement of the kinetic energy and of the direction of the incoming particles. Therefore, a deep learning based event reconstruction chain has been designed to reconstruct these important physics observables. The choice is motivated by the fact that deep learning models are very effective when working with particle detectors in which a variety of electrical signals are produced and may be treated as low-level features. The DL-based event reconstruction is trained on dedicated Monte Carlo simulation and tested on both simulated and test-beam data. All research activities concerning the Limadou HEPD-02 event reconstruction are carried out exploiting the INFN computing infrastructure: all the steps of the MC simulation (Geant4 + signal digitization) and the data pre-processing are carried out exploiting parallel submission on CNAF nodes, on the other hand the training of Deep Learning models is performed on GPUs available at RECAS.
In this contribution, the HEPD-02 deep learning based event reconstruction be described as well as its computing strategy based on INFN available resources.
