Track identification in Phase II of Project 8 with machine learning techniques

Project 8 is a direct neutrino mass experiment using the tritium endpoint method. To achieve the sensitivity goal of 40 meV, the Project 8 collaboration has developed the technique of Cyclotron Radiation Emission Spectroscopy (CRES) to measure the energy of a beta electron from the frequency of its cyclotron radiation in a magnetic field. In Phase II of Project 8, the electron signals show up as tracks on top of the noise background in spectrograms. In this talk, I will present our progress in using machine learning methods to extract electron signals from the noise background in simulated Phase II data.