21 January 2021 - Online event

Abstract

Machine learning and deep learning algorithms have gained importance in particle/astroparticle physics in the last years. They play a key role in the event reconstruction of the most modern experiments: from the identification of the particle nature and tracking, up to energy reconstruction and detection of data anomalies.
The attractive feature of these techniques is their ability to model large dimensionality inputs and exploit not trivial correlations among the variables, which could be hidden or not easy to model. The key of the learning process is in the training on Monte Carlo samples. For this reason experimental collaborations are investing resources to realize dedicated simulations as representative as possible of real data.
This talk focuses on the applications of deep learning tools in the CSES mission, illustrating the strategy and the performances of a new event reconstruction based on neural networks for the Limadou High Energy Particle Detector.