The ARIANNA experiment is a proposed Askaryan detector designed to record radio signals induced by neutrino interactions in the Antarctic ice. Because of the low neutrino flux at high energies, the physics output is limited by statistics. Hence, an increase in sensitivity will significantly improve the interpretation of data and will allow us to probe new parameter spaces. The trigger thresholds are limited by the rate of triggering on unavoidable thermal noise fluctuations. Here, we present a real-time thermal noise rejection algorithm that will allow us to lower the thresholds substantially and increase the sensitivity by up to a factor of two compared to the current ARIANNA capabilities. A deep learning discriminator, based on a Convolutional Neural Network (CNN), was implemented to identify and remove a high percentage of thermal events in real time while retaining most of the neutrino signal. We describe a CNN that runs efficiently on the current ARIANNA microcomputer and retains 94% of the neutrino signal at a thermal rejection factor of $10^5$. Finally, we report on the experimental verification from lab measurements.
Askaryan; UHE neutrinos; in-ice radio detection; trigger optimization; radio; deep learning
|Subcategory||Experimental Methods & Instrumentation|
|Collaboration||other (fill field below)|