Faceted Eyes on the Universe: Event Reconstruction in IceCube Upgrade Using Transformer-based Neural Networks for Multi-PMT mDOM Optical Sensors

• Applicant: Professor Dr. Christopher Wiebusch
• Subject Area: Nuclear and Elementary Particle Physics, Quantum Mechanics, Relativity, Fields ; Astrophysics and Astronomy
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 551395962

The IceCube Neutrino Observatory, through its upcoming Upgrade, is targeting cutting-edge neutrino oscillation measurements. The Upgrade will significantly enhance the observatory's capabilities by incorporating 700 new optical sensors, including 400 multi-PMT digital optical modules (mDOMs), into its existing array. These enhancements are aimed at improving the detection and analysis of atmospheric neutrino events down to a few GeV energy. The mDOM as a pivotal component of the Upgrade features a strongly enlarged, omnidirectional and segmented photo-detection area. This poses challenges to existing reconstruction methods that could not be fully resolved to date. We propose the application of transformer-based neural network algorithms for event reconstruction, which will address the increased complexity and volume of data expected from the enhanced sensor array. By leveraging state-of-the-art machine learning techniques, specifically transformer neural networks, the project seeks to refine the reconstruction of neutrino interactions within the ice, enhancing the resolution of event characteristics such as direction and energy. This initiative not only promises to advance our understanding of neutrino properties and interactions but also strengthens the collaborative scientific efforts between South Korea and Germany.

• DFG Programme: Research Grants
• International Connection: South Korea
• Cooperation Partner: Professor Dr. Chang Hyon Ha