Investigation Into Improving CUORE Energy Reconstruction Using Principal Component Analysis

The Cryogenic Underground Observatory for Rare Events (CUORE), located in the Laboratori Nazionali del Gran Sasso in Italy, is an experiment searching for the neutrinoless double-beta decay (0νββ) of 130Te. Observing neutrinoless double-beta decay would have a number of implications, including indicating that neutrinos are their own antiparticle. This would in turn indicate that Lepton number is not conserved and help explain why there is more matter than antimatter in the known universe. In this talk, we present a first investigation into the use of machine learning models and in particular, Principal Component Analysis, to improve energy reconstruction for the CUORE experiment. Sensitivity to 0νββ in the background dominated regime is limited by fluctuations in the background and signal energy resolution. Excellent energy resolution is critical for CUORE as it helps to more clearly distinguish 0νββ signals from background events.The current method of energy reconstruction, optimal filter, assumes that noise is stationary (i.e. does not change during the pulse) and that signals from different energies have the same pulse shape (linear response), which we know is not true for the CUORE experiment. Thus, taking into account these factors using PCA could lead to an improved model for energy reconstruction. In addition, the techniques we are developing could be employed in the next generation experiment CUPID to help reach an unprecedented sensitivity to 0νββ.