“Baryon Number Violation in DUNE: Proton Decay Sensitivity using Machine Learning and LArTPC Validations with ProtoDUNE-2”
The Deep Underground Neutrino Experiment (DUNE) is a long-baseline neutrino oscillation experiment at Fermilab. Its primary goals are to resolve the neutrino mass hierarchy and measure the charge-parity (CP) violating phase in the lepton sector, an indicator of possible explanations for our matter-dominated universe. In addition to these aims, the DUNE physics program will include baryon number-violating searches, such as proton decay. The experiment’s liquid argon time projection chamber (LArTPC) far detectors have a total volume of 70kton and are located approximately 1500m underground.
These LArTPCs will provide excellent imaging capabilities, as seen by the ProtoDUNE detectors, for detecting rare events. This talk will begin by covering the ongoing validation of the DUNE LArTPC detector technology at the CERN Neutrino Platform, with a primary focus on the ProtoDUNE-2 experiment. In the second part, I will present a machine learning-based analysis for DUNE’s sensitivity to the proton decay channel p → K+ν with the DUNE Far Detectors. This analysis utilizes NuGraph, a graphical neural network that operates directly on detector hits, and a Boosted Decision Tree (BDTs) to enhance event classification. These approaches demonstrate the potential of machine learning in application to LArTPCs such as DUNE in rare process searches.
Host: Karsten Heeger