Hybrid-graph neural network method for muon fast reconstruction in neutrino telescopes

TL;DR

This paper presents a hybrid-graph neural network approach for rapid and accurate muon track reconstruction in neutrino telescopes, significantly outperforming traditional methods in speed while maintaining high precision.

Contribution

Introduction of a novel hybrid-graph neural network method that accelerates muon reconstruction and includes a quality assessment mechanism for neutrino telescope data analysis.

Findings

Achieves 0.19-0.29 ms per event GPU runtime, 3 orders faster than traditional methods.

Median angular error of ~0.1° for high-energy muons (10-100 TeV).

Reconstructed Cherenkov photon positions within 3-5 meters.

Abstract

Fast and accurate muon reconstruction is crucial for neutrino telescopes to improve experimental sensitivity and enable online triggering. This paper introduces a hybrid-graph neural network (GNN) method tailored for efficient muon track reconstruction, leveraging the robustness of GNNs, alongside traditional physics-based approaches. The "light GNN model" achieves a run-time of 0.19-0.29 ms per event on GPUs, offering a 3 orders of magnitude speedup compared to traditional likelihood-based methods, while maintaining a high reconstruction accuracy. For high-energy muons (10-100 TeV), the median angular error is approximately 0.1{\deg}, with errors in reconstructed Cherenkov photon emission positions being below 3-5 m, depending on the GNN model used. Furthermore, the semi-GNN method offers a mechanism to assess the quality of event reconstruction, enabling the identification and exclusion of poorly reconstructed events. These results establish the GNN-based approach as a promising solution for next-generation neutrino telescope data reconstruction.