This is a project performed by Florian Führer, Tom Charnock and Anne Zilles.
Based on ZHAireS simulations for Cosmic Rays and neutrino events for a GRAND-like toymodel array, we studied the efficiency of using Deep Neural Networks to identify signals in a background dominated environment. Those detected signals should then feed neural networks which are able to do a fast online-reconstruction of the shower parameters based on the time traces collected from the array. (for the current status see eg. GRAND wiki:http://www.iap.fr/grand/wikigrand/index.php?title=File:Fuehrer_TriggerNN_CallDec2019.pdf)
Temporary results can be found in the ARENA 2018 proceedings: https://arxiv.org/pdf/1809.01934.pdf