Developing a Bubble Chamber Particle Discriminator Using Semi-Supervised Learning

The identification of non-signal events is a major hurdle to overcome for bubble chamberdark matter experiments such as PICO-60. The current practice of manually developing a discriminator function to eliminate background events is difficult when available calibration data is frequently impure and present only in small quantities. In this study, several different discriminator input/preprocessing formats and neural network architecturesare applied to the task. First, they are optimized in a supervised learningcontext. Next, two novel semi-supervised learningalgorithms are trained, and found to replicate the Acoustic Parameter (AP) discriminator previously used in PICO-60 with a mean of 97% accuracy.