The Jiangmen Underground Neutrino Observation (JUNO) experiment is designed to measure the neutrino mass order (NMO), one of the biggest remaining puzzles in neutrino physics. Regarding the sensitivity of JUNO’s NMO measurement, besides the precise measurement of reactor neutrino‘s energy spectrum as the primary source, the independent measurement of atmospheric neutrino oscillation has a great potential to enhance the sensitivity in the combined analysis. The NMO sensitivity from the atmospheric neutrino measurement heavily relies on the angular resolution of the reconstructed incident angle of the neutrino. Thus, a powerful direction reconstruction algorithm is urgently needed in JUNO.
This poster introduces a novel method for directionality reconstruction, by extracting multiple features from the PMT waveform and using them as the input of machine learning models, including a deep convolutional neural network and a spherical graph neural network. The preliminary results based on JUNO simulation show that this method has great potential to reconstruct the incident angle precisely, for both atmospheric neutrino and charged leptons. In this poster, the machine learning method and the performance estimation will be briefly presented.