Enhancing Events in Neutrino Telescopes through Deep Learning-Driven Super-Resolution

📅 2024-08-16
🏛️ arXiv.org
📈 Citations: 0
Influential: 0
📄 PDF

career value

188K/year
🤖 AI Summary
Neutrino telescopes suffer from sparse photon sampling due to large optical module spacing, severely limiting angular resolution in muon event reconstruction. To address this, we propose the first event super-resolution method explicitly designed for realistic detector geometries: physically consistent insertion of virtual optical modules within ice or water media, jointly modeling photon transport physics and detector response; and a lightweight convolutional network that directly reconstructs high-resolution photon distributions from sparse hit data—fully compatible with both conventional and deep-learning-based reconstruction pipelines. Experiments demonstrate significant improvement in angular resolution for muon events in ice-based telescopes such as IceCube (average gain ~15%), with seamless transferability to water-based detectors like KM3NeT and diverse event types. This work establishes a novel paradigm for high-precision particle reconstruction under sparse-sampling conditions.

Technology Category

Application Category

📝 Abstract
Recent discoveries by neutrino telescopes, such as the IceCube Neutrino Observatory, relied extensively on machine learning (ML) tools to infer physical quantities from the raw photon hits detected. Neutrino telescope reconstruction algorithms are limited by the sparse sampling of photons by the optical modules due to the relatively large spacing ($10-100,{ m m})$ between them. In this letter, we propose a novel technique that learns photon transport through the detector medium through the use of deep learning-driven super-resolution of data events. These ``improved'' events can then be reconstructed using traditional or ML techniques, resulting in improved resolution. Our strategy arranges additional ``virtual'' optical modules within an existing detector geometry and trains a convolutional neural network to predict the hits on these virtual optical modules. We show that this technique improves the angular reconstruction of muons in a generic ice-based neutrino telescope. Our results readily extend to water-based neutrino telescopes and other event morphologies.
Problem

Research questions and friction points this paper is trying to address.

Neutrino Telescope
Light Detection Limitations
Muon Direction Accuracy
Innovation

Methods, ideas, or system contributions that make the work stand out.

Deep Learning
Virtual Sensors
Neutrino Telescopes
F
Felix J. Yu
1The NSF AI Institute for Artificial Intelligence and Fundamental Interactions, 2Department of Physics and Laboratory for Particle Physics and Cosmology, Harvard University, Cambridge, MA 02138, US
N
Nicholas Kamp
2Department of Physics and Laboratory for Particle Physics and Cosmology, Harvard University, Cambridge, MA 02138, US
C
Carlos A. Argüelles
1The NSF AI Institute for Artificial Intelligence and Fundamental Interactions, 2Department of Physics and Laboratory for Particle Physics and Cosmology, Harvard University, Cambridge, MA 02138, US