7th Inter-Experimental LHC Machine Learning Workshop

Electron-Proton Classification with AMS ECAL using Convolutional Vision Transformers

Not scheduled

20m

61/1-201 - Pas perdus - Not a meeting room - (CERN)

Poster 1 ML for object identification and reconstruction Poster Session

Speaker

Berk Turk (Middle East Technical University (TR))

Description

Alpha Magnetic Spectrometer (AMS-02) is a precision high-energy cosmic-ray experiment on the ISS operating since 2011 and has collected more than 240 billion cosmic ray events. Among them, positrons are important in understanding the particle nature of dark matter. Classifying the positron signals is challenging due to the abundant background of cosmic ray protons. Therefore, we use a state-of-the-art convolutional and transformer model, CoAtNet, that employs the shower signals from the ECAL to classify the electrons/positrons in the dominant cosmic proton background. We created AMS ECAL shower datasets of purer samples of electrons and protons from ISS data and test beam data in the 50-200 GeV reconstructed energy range by applying cuts. We extend the energy range of the ISS datasets up to 1 TeV by building a charge-confused proton and electron classifier with a transformer using reconstructed variables from Tracker, Time of Flight, and Ring Imaging Cherenkov Detector. We trained and fine-tuned the CoAtNet ECAL model using the ISS datasets in the 50-500 GeV range and obtained a proton rejection that was greater than the traditional methods by a factor of 2 on ISS Data.