Improving sensitivity of the ARIANNA detector by rejecting thermal noise with deep learning

Abstract

The ARIANNA experiment is an Askaryan detector designed to record radio signals induced by neutrino interactions in the Antarctic ice. Because of the low neutrino flux at high energies (Eν > 1016 eV), the physics output is limited by statistics. Hence, an increase in sensitivity significantly improves the interpretation of data and offers the ability to probe new parameter spaces. The amplitudes of the trigger threshold are limited by the rate of triggering on unavoidable thermal noise fluctuations. We present a real-time thermal noise rejection algorithm that enables the trigger thresholds to be lowered, which increases the sensitivity to neutrinos by up to a factor of two (depending on energy) compared to the current ARIANNA capabilities. A deep learning discriminator, based on a Convolutional Neural Network (CNN), is implemented to identify and remove thermal events in real time. We describe a CNN trained on MC data that runs on the current ARIANNA microcomputer and retains 95% of the neutrino signal at a thermal noise rejection factor of 105, compared to a template matching procedure which reaches only 102 for the same signal efficiency. Then the results are verified in a lab measurement by feeding in generated neutrino-like signal pulses and thermal noise directly into the ARIANNA data acquisition system. Lastly, the same CNN is used to classify cosmic-rays events to make sure they are not rejected. The network classified 102 out of 104 cosmic-ray events as signal.

Tau neutrinos in the next decade: from GeV to EeV

“…ARIANNA derived a limit on the high-energy neutrino flux which demonstrates the feasibility of the in-ice radio detection technique [646]. The ARIANNA detector is also a test bench for future detector optimizations relevant for the future IceCube-Gen2, e.g., the detector was optimized through optimizations of the signal chain [647] and trigger [646]. Furthermore, reconstruction algorithms for the neutrino energy, direction and flavor were developed, and probed with in situ measurement using radio emitters that are lowered into the ice, as well as through the measurement of cosmic rays [454,[648][649][650][651][652][653][654][655][656][657][658].…”

“…Additional upward pointing LPDAs were added for a cosmic-ray detection and vetoing of anthropogenic noise. ARIANNA derived a limit on the high-energy neutrino flux which demonstrates the feasibility of the in-ice radio detection technique [646]. The ARIANNA detector is also a test bench for future detector optimizations relevant for the future IceCube-Gen2, e.g., the detector was optimized through optimizations of the signal chain [647] and trigger [646].…”

Results from the ARIANNA high-energy neutrino detector

“…The trigger threshold is then set by the maximum data rate the detector can handle. To get around these limitations, we developed a real-time thermal noise rejection algorithm that enables the trigger thresholds to be lowered, which increases the sensitivity to neutrinos by up to a factor of two (depending on energy) compared to the current ARIANNA capabilities [16]. A deep learning discriminator, based on a Convolutional Neural Network (CNN), was implemented to identify and remove thermal events in real-time.…”

Developing New Analysis Tools for Near Surface Radio-based Neutrino Detectors

“…In contrast to the more traditional techniques used to develop the previous selection criteria, a new cut was developed using a deep learning approach which exploits the unique "chirped" signal from the LPDA. A convolutional neural network (CNN) architecture was chosen, which builds on previous work [11] in which CNNs were found to be more efficient at discriminating between neutrino signal and noise compared to fully connected neural networks.…”