Date Thesis Awarded


Access Type

Honors Thesis -- Access Restricted On-Campus Only

Degree Name

Bachelors of Science (BS)




David Armstrong

Committee Members

Seth Aubin

Pierre Clare


This report presents the development of a machine learning algorithm (a neural net- work) for the purpose of track reconstruction in the MOLLER experiment. The MOLLER experiment is a collaboration at Jefferson Lab which plans on measuring the parity-violating asymmetry in high energy electron-electron (Møller) scattering. This measurement is an important test of the Standard Model and could potentially serve as evidence for new physics beyond the Standard Model.

Reconstruction of electron trajectories provides important kinematic data about the electrons which is necessary for the asymmetry measurement. As such, the MOLLER experiment requires a track reconstruction tool which can efficiently handle large amounts of data. We created a recurrent neural network which models the trajectories of electrons in the detector system for MOLLER. We trained the network using data from the GEANT 4 Monte Carlo simulation for the experiment.

At this stage, our network is able to correctly predict the locations where an electron hits the main detector given that electron’s previous positions in the detector system. We describe various studies we conducted to improve the accuracy and efficiency of our network, as well as present suggestions for any potential continuations of this project.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

On-Campus Access Only