A convolutional neural network based cascade reconstruction for the IceCube Neutrino Observatory

Abstract

A: Continued improvements on existing reconstruction methods are vital to the success of high-energy physics experiments, such as the IceCube Neutrino Observatory. In IceCube, further challenges arise as the detector is situated at the geographic South Pole where computational resources are limited. However, to perform real-time analyses and to issue alerts to telescopes around the world, powerful and fast reconstruction methods are desired. Deep neural networks can be extremely powerful, and their usage is computationally inexpensive once the networks are trained. These characteristics make a deep learning-based approach an excellent candidate for the application in IceCube. A reconstruction method based on convolutional architectures and hexagonally shaped kernels is presented. The presented method is robust towards systematic uncertainties in the simulation and has been tested on experimental data. In comparison to standard reconstruction methods in IceCube, it can improve upon the reconstruction accuracy, while reducing the time necessary to run the reconstruction by two to three orders of magnitude.

Ultra-High-Energy Cosmic Rays: The Intersection of the Cosmic and Energy Frontiers

“…In recent years, first applications were developed to adapt machine learning-based analysis techniques in MMA, including gamma-ray astronomy [707], neutrino astrophysics [708], gravitationalwave detection [709], and, as seen in Sec. 5.3.3 through Sec.…”

High-Energy and Ultra-High-Energy Neutrinos

“…18, will soon open a new channel for multi-messenger analyses [599]. New event reconstructions relying on updated likelihood constructions [613] or advanced convolutional neural networks [614] show promise in improving the angular and energy resolution necessary for precise characterization of astrophysical fluxes. Recent updates in the point-source likelihood used by IceCube searches aim to improve modeling by utilizing a fuller description of its point spread function [615].…”

Graph Neural Networks and Application for Cosmic-Ray Analysis

“…The simplest possible solution to use a full shower footprint is by using CNN, as shown in [29] for IceCube. The input pixels of the CNN is the individual photo-detector with electronics, i.e., Digital Optical Modules (DOMs) of the detector.…”

“…3, the detector will shift to a more irregular geometry. Data transformation of such future detector configuration to an orthogonal grid and specialized kernels (as shown in [29]) will possibly be complex and inefficient. As described in earlier sections, GNNs surely hold promise in their capability to allay the limits of MLPs and CNNs.…”