Particle Identification in the Muon Forward Tracker using Machine Learning

In this work, the particle identification capabilities of the ALICE Muon Forward Tracker (MFT) are explored. The goal is to detect (anti-)nuclei at forward rapidity ($-3.6 < \eta < -2.4$) for the first time. This measurement is crucial to improve rapidity dependent uncertainties in nucleosynthesis models. The study presented in this note is performed on the data collected by the ALICE experiment during RUN 3. A Machine learning algorithm is developed to optimise the separation of nuclei and Minimum Ionising Particles (MIP) in the MFT using anomaly detection. The resulting statistical significance for nuclei is $> 30$ which means that the MFT can be used to detect (anti-)nuclei at forward rapidity. In combination with the high charge-reconstruction efficiency of the MFT for $p_{\text{T}} < 2\,$GeV, anti-nuclei could be distinguished.