Speaker
Steven Barwick
(UC Irvine)
Description
Located atop of thick ice sheets, radio-based high energy neutrino detectors search for astrophysical neutrinos with energies greater than $10^{16}$ eV. Cosmic rays at similar energies that interact in the atmosphere provide both a calibration source (due to pulsed radio emission similar to neutrinos) and background. I will discuss a recently implemented real time trigger based on convolutional neural nets (CNNs) which offer the potential to increase the neutrino effective volume by factors 1.5 to 3, depending on neutrino energy, relative to the most commonly assumed trigger that simply require unusual excursions from random noise fluctuations. The CNN trigger concept should also improve the sensitivity to atmospheric cosmic rays.