… In this work, we leverage graph neural networks to improve signal detection in the IceCube
… application of graph neural networks to the challenging problem of signal detection in the …

… of signal photons produced per event. To address this challenge, it is possible to … IceCube
events as point cloud graphs and use a Graph Neural Network (GNN) as the classification and …

… in IceCube, but only for signal classification [10]. The approach presented in this proceeding
is among the first to employ GNNs for reconstruction and interaction type classification. …

… For the neutrinos detectable by IceCube, the signal in a neutrino event may range from a low
of a few to a high of order 100,000 detected photoelectrons. In the energy range this work is …

… In MC data, this work finds a 15% increase in neutrino signal as compared to … graph neural
networks (GNN’s) have been shown to work well with irregularly spaced data in classification …

… reconstruction and particle classification. The algorithms … the versatile framework of graph
neural networks (GNNs), we … structure, with each neuron sending signals along to the next …

… of using such methods at the IceCube Neutrino Observatory. … This can be easily used to
either do node classification or … further in the ice to deposit signals in the IceCube (IC) array (…

… In this thesis, I present two Graph Neural Networks (GNNs) … that signal and then send it up
to the IceCube laboratory on the … filter as a neural network is a standard classification task, one …

… In conclusion, the dynedge is a reconstruction approach based on Graph Neural Networks
(GNN) customized for IceCube events. By using a low-energy simulated dataset used in …

… The DeepCore sub-detector of the IceCube Neutrino Observatory provides access to … on a
convolutional neural network is applied, providing increased signal efficiency and background …