Machine Learning for MuSIC@Indiana Event Identification

The Multi-Sampling Ioniozation Chamber (MuSIC@Indiana) is a frisch-grided ionization chamber than can effectively and efficiently measure fusion cross sections at near-barrier energies [1]. Previously, this detector was used to measure the fusion cross section of ¹²C with beams of ^16-20O [2]. In order to determine the cross section, the evaporation residues of the reaction must be separated from other events that occur in the detector. We aim to leverage machine learning techniques to quickly and efficiently separate different classes of events. This will be done utilizing a polynomial support vector machine (SVM). Furthermore, we are developing a machine learning model trained on simulated data, which will allow us to have an online analysis available for future experiments. The method will be employed on an upcoming ¹⁴O+¹²C experiment, the results of which will be presented in this talk.

This work is supported by DE-SC0022299 - TRAIN-MI Program: High Energy Physics Instrumentation Traineeship in Michigan and NSF PHY-2309923 WoU-MMA: Studying neutron stars through the lens of nuclear reactions.

[1] J. Johnstone et al., Nucl. Instr. Meth. A 1014, 166697 (2021).

[2] Hudan, Sylvie, et al. Physical Review C. 109. 10.1103 (2024).