Rare event searches allow us to search for new physics at energy scales inaccessible with other means by leveraging specialized large-mass detectors. Machine learning provides a new tool to maximize the information provided by these detectors. The information is sparse, which forces these algorithms to start from the lowest level data and exploit all symmetries in the detector to produce results. In this work we present KamNet which harnesses breakthroughs in geometric deep learning and spatiotemporal data analysis to maximize the physics reach of KamLAND-Zen, a kiloton scale spherical liquid scintillator detector searching for neutrinoless double beta decay (0νββ). With the help of KamNet, KamLAND-Zen provides a limit that reaches below 50 meV for the first time and is the first search for 0νββ in the inverted mass ordering region. A key component of this work is the addition of an attention mechanism to elucidate the underlying physics KamNet is using for the background rejection.
Bio: Aobo Li received his B.S. in physics at the University of Washington in 2015, then went on to do his graduate work at Boston University. He did his graduate work as part of the KamLAND-Zen collaboration, searching for neutrinoless double-beta decay with monolithic liquid scintillator detectors. After getting his Ph.D. in 2020, Aobo joined UNC Chapel Hill as a Postdoctoral Research Associate and COSMS Fellow. Aobo initiate and lead the Ge Machine Learning (GeM) group, bringing AI solutions to two Ge detector experiments: Majorana Demonstrator and LEGEND. He is also one of the recepient of APS Dissertation Award in Nuclear Physics this year.
Ge detector Interpretable Machine Learning: https://inspirehep.net/literature/2121002