TES X-ray pulse identification using CNNs

Póster presentado a la XXX Astronomical Data Analysis Software and Systems Conference, celebrada online del 8 al 12 de noviembre de 2020 en la University of Maryland, College Park Campus, USA.[Aims]: to detect and to reconstruct the pulses that X-ray photons produce at TES (Transition Edge Sensor) detectors (like the one in the Athena/X-IFU instrument), using Deep Learning techniques. The dataset used contains simulations performed with the Athena official simulator SIXTE.[Methods]: We construct and train a Convolution Neural Network (CNN) to differentiate between single, double and triple pulses. We use a hyper-parameter bayesian optimization to select the optimal CNN architecture.[Results]: we present the results of our CNN classification over 15,000 simulated pulses achieving excellent performance metrics.AHEAD (Integrated Activities for the High Energy Astrophysics Domain).Peer reviewe