Undergraduate Research at Jefferson Lab

The Configuration and Teaching of Hydra for use in Hall A's Research

Student: Jacob Dulya

School: Duquesne University

Mentored By: Alexandre Camsonne

The quality of data generated in an experiment can be negatively impacted by hardware problems or maintenance mistakes. This can be seen through data taken before the error was remedied and therefore can be seen in graphs made with said data. Human analysts can be slow and can miss discrepancies indicative of issues, Hydra aims to fix that. Hydra is a way that graph images can be monitored to spot data issues before they persist as well as contribute to data analysis. Hydra is software developed with machine learning technology designed to analyze graph images input from Jefferson Lab. Hydra was trained on images of plotted data, and can be used alongside research to detect issues with generated data. It is important to note that the original plan for the project included more graphs and training for Hydra beyond the proof of concept that was achieved, but even so sufficient work has been done to show Hydra's capability to be integrated to Hall A.

[Watch the presentation on YouTube]