Deep Learning techniques for signal processing and event reconstruction in DUNE

DUNE - Deep Underground Neutrino Experiment - will be an experiment based in the USA whose main goal will be to study long-baseline neutrino oscillations from an accelerator beam. At CERN, a smaller prototype of the DUNE far detectors has been created, ProtoDUNE, with the objective to test and validate the technology required for DUNE far detectors. The signal captured by ProtoDUNE needs to be processed to clean it from noise. To do so, a Deep Learning tool has been developed at CERN. However, the costs of this tool imply the necessity to evaluate its performance comparing to more classical techniques, which involve lower costs. For this reason, a statistical algorithm has been considered, which is the one currently used: the results of the two solutions are compared in this project.