Deep Learning In Neutrino Physics

The Department of Physics in the College of Arts and Sciences is honored to welcome Dr. Fernanda Psihas to present their weekly virtual colloquia. Dr. Psihas is a Research Associate at Fermilab working on Neutrino Experiments and Deep Learning.

Abstract: Among all fundamental particles, neutrinos are the least understood. Experiments worldwide study their properties and behavior, which could be linked to questions about the formation of our Universe. Over the past several years, physicists have adapted techniques from computer vision, whose tasks translate naturally into many of those needed for detector data analysis.

The adaptation of deep learning for particle physics has been fruitful, and is becoming more widespread. These techniques have yielded improvements to the reach of many experiments and redefined the limit to what is attainable for data collection, analysis, and R&D. Deep learning has also opened the door to new opportunities to evaluate uncertainties associated with our simulations of particle behavior.

In this overview of deep learning for neutrino experiments, Dr. Psihas will showcase applications to detector data analysis developed in the past few years, including techniques in development for future experiments, and discuss the challenges associated with the use of Deep Learning in physics.