KM3NeT is a research infrastructure hosting two large-volume Cherenkov neutrino detectors which are currently under construction in the Mediterranean Sea. The KM3NeT/ARCA detector is optimised for the detection of high-energy neutrinos from astrophysical sources in the TeV-PeV energy range. Once completed, the detector will consist of 230 detection units. Here, we present a Deep Learning method using graph neural networks that is trained and applied to events gathered with 6 and 8 active detection units of KM3NeT/ARCA. Graph neural networks have been trained for classification and regression tasks, showing very promising performances in a range of different tasks like neutrino-background identification, neutrino event topology classification, energy and direction reconstruction, and also in the study of properties of muon bundles.
Data reconstruction and classification with graph neural networks in KM3NeT/ARCA6-8
Guidi C.;Malerba L.;Marconi M.;Parisi V.;Romanov A.;Sanguineti M.;Taiuti M.;Vannoye G.;
2024-01-01
Abstract
KM3NeT is a research infrastructure hosting two large-volume Cherenkov neutrino detectors which are currently under construction in the Mediterranean Sea. The KM3NeT/ARCA detector is optimised for the detection of high-energy neutrinos from astrophysical sources in the TeV-PeV energy range. Once completed, the detector will consist of 230 detection units. Here, we present a Deep Learning method using graph neural networks that is trained and applied to events gathered with 6 and 8 active detection units of KM3NeT/ARCA. Graph neural networks have been trained for classification and regression tasks, showing very promising performances in a range of different tasks like neutrino-background identification, neutrino event topology classification, energy and direction reconstruction, and also in the study of properties of muon bundles.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.



