CUPID (CUORE Upgrade with Particle IDentification) is a foreseen tonne-scale array of (LMO) cryogenic calorimeters with double readout of heat and light signals. One of the dominant background sources is pile-up of standard-model double beta decay events in the same crystal detector.
It is crucial to achieve a time difference pile-up rejection of ~ 300 us while keeping a high efficiency for single decay events.
The goal of this project is to develop machine learning techniques to perform event classification with supervised learning of heater generated pile-up events.
CUORE will conclude data taking in ~ 3 years from now, and pile-up rejection is an issue that needs to be timely addressed to continue the experimental program of neutrino-less double beta decay search with CUPID.
Machine learning techniques have been very recently included in the data processing of cryogenic bolometer data with outstanding results. The CUORE experiment has replaced its pulse shape analysis algorithm based on the optimum filter (PSA) with a principal component analysis (PCA) obtaining an improvement both in the global efficiency as in the energy threshold. The discrimination of signal-like pulses from spurious triggers would suffer of a very low efficiency
The current ability to discriminate pile-up in CUPID-like bolometers needs improvement by an order of magnitude in time difference resolution [2,3] to guarantee the required background goal. To overcome this bound it is necessary to explore new techniques with the potential of fully extracting the information in the available data sample.
Improving the pile-up rejection time resolution will allow to fully explore the inverted hierarchy region of neutrino masses and to push knowledge about the Majorana nature of neutrinos further.
[1] Adams, D. Q. et al. (CUORE Collaboration), High sensitivity neutrinoless double-beta decay search with one tonne-year of CUORE data, [arXiv:2104.06906]
[2] The CUPID Interest Group, A novel technique for the study of pile-up events in cryogenic bolometers, [arXiv:2011.11726]
[3] The CUPID Interest Group, CUPID pre-CDR, [arXiv:1907.09376]