EXO-200 has published its first paper on energy and position reconstruction with Deep Learning. We demonstrate that convolutional neural networks can reach and exceed conventional method of energy and position reconstruction. The PI initiated and led the Deep Learning effort on EXO-200.
NVIDIA granted our group a TitanX GPU to pursue Deep Learning techniques for the EXO-200 data analysis. For certain applications, these powerful machines can outperform conventional CPUs by an order of magnitude. We will use the TitanX GPU to train convolutional neural network event classifiers and to track photons in EXO-200. (As of 2020, we have a total 6 TitanXP GPUs in the group available thanks to support by NVIDIA and DOE)
Majoron is a hypothetical particle associated with spontaneous lepton number symmetry breaking. Majorons could mediate neutrinoless double beta decay, which produce a specific signature in the EXO-200 data. Maximum likelihood shape analysis of the EXO-200 data found no evidence of the Majoron-mediated decays. The results of this study were included in the Particle Data Group annual review. From PRD90 (2014) 092004
Comparison between the theoretically predicted (red line) and experimentally observed (black points) spectral shape of the beta‐decay of Xe‐137 to the ground state of Cs‐137. This work is published in Phys. Rev. Lett. 124, 232502, 2020
A scanning electron microscope image of a silicon photomultiplier (SiPM) studied in our lab. Our paper on the first measurement of an angular-resolved reflectivity and photodetector efficiency of a SiPM in liquid xenon is published in JINST 15 (2020) P01019