Application of Machine Learning with the Minimum Bias Detector (MBD) in sPHENIX

The PHENIX detector is one of two main detectors used to track high-energy collisions at the Relativistic Heavy Ion Collider (RHIC) at Brookhaven National Laboratory. Designed to collect nuclear data for analysis, these detectors have expanded our collective knowledge of quark-gluon plasma (QGP). An upgrade to PHENIX, called sPHENIX, will enable far better measurements of upsilon production and heavy flavored jets. The minimum bias detector (MBD) is a subsystem of sPHENIX that acts as the primary trigger for collisions and uses the original beam-beam counter (BBC) from PHENIX. Instead of using standard digital signal processing techniques to extract the time of arrival and charge in the MBD, I used machine learning techniques such as boosted-decision trees, convolutional neural networks, and linear regression models to obtain these values. I have built various models utilizing these techniques to compare the accuracies and mean-average errors (MAE), as well as the speed of execution. At the beginning of sPHENIX data-taking this program will be used in the reconstruction of the MBD data.