Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 82: Neutrino Astronomy IV

T 82.6: Vortrag

Donnerstag, 19. März 2026, 17:30–17:45, KS H C

Machine Learning for DSNB Detection in JUNO: Utilizing Spatial-Temporal PMT Hit Patterns for Background Suppression — •David Maksimovic¹, Michael Wurm¹, Daniel Tobias Schmid¹, and Dhaval J. Ajana² — ¹Johannes Gutenberg-University — ²Florida State University

The detection of the Diffuse Supernova Neutrino Background (DSNB) poses a significant challenge in neutrino astronomy, primarily due to backgrounds mimicking the extremely rare antineutrino events via Inverse Beta Decay (IBD) . The Jiangmen Underground Neutrino Observatory (JUNO) uses a liquid scintillator to detect these neutrinos in the 12 to 30 MeV range. There, especially Neutral-Current (NC) interactions of atmospheric neutrinos dominate the predicted DSNB signal by 1-2 orders of magnitude. In this talk, we present a comparative study of ML discrimination algorithms, ranging from 3D Convolutional Neural Networks (3D CNNs), LSTMs and Convolutional LSTMs (ConvLSTMS) and utilizing Fourier Transformations. These techniques analyze time-sequenced data from photomultiplier tube (PMT) hit patterns to capture the spatial-temporal dynamics of particle interactions. Here we present the resulting background reduction capabilities for JUNO.

Keywords: Machine Learning; Diffuse Supernova Neutrino Background; JUNO; Data preparation