93 research outputs found
Deep Neural Networks for energy reconstruction of Inverse Beta Decay events in JUNO
No full textThe Jiangmen Underground Neutrino Observatory (JUNO) is a scintillation detector, currently under construction, which aims to solve the neutrino mass hierarchy by measuring reactor electron antineutrino energy spectrum with a a resolution of 3%/sqrt(E(MeV)) – the highest ever achieved in a large mass neutrino detector. Several approaches for energy reconstruction are being evaluated on simulated data, and Deep Learning methods have already shown promising results, both in accuracy and in efficiency. In this work, a new Convolutional Neural Network with a rotational invariant architecture is trained on a small dataset of 160k instances, and is fine-tuned to exploit the detector’s spherical symmetry and make use of position and timing data from individual photomultipliers. This approach proves to be insensitive to the presence of dark noise from thermal fluctuations, leading to a (2.45+-0.03)% visual energy resolution at 2 MeV, only slightly higher than the 2.2% expected from theory, with a reconstruction bias well below 1%. However, a simpler Fully Connected Neural Network, replicated from previous work, which uses only integral data and is trained on a larger dataset (750k instances), leads to a slightly better resolution of (2.26+-0.05)% at 2 MeV, while being more sensitive to added noise – proving that there could still be some margin of improvement for more complex methods.ope
Deep Neural Networks per la ricostruzione dell’energia di eventi di decadimento beta inverso nell’esperimento JUNO
No full textModern experiments on the intensity frontier requires complicate algorithms to extract data for physical analysis. The neural
network techniques, having received a considerable boost during the last years, are becoming a useful tool for addressing many tasks
of data processing and provide in some cases better performance than traditional methods. The Jiangmen Underground Neutrino
Observatory (JUNO), a next generation experiment under construction in South of China, has been designed to measure the neutrino
mass hierarchy. Moreover, thanks to its large active mass, JUNO will be able to observe neutrinos coming from different sources:
solar neutrinos, atmospheric neutrinos, geo-neutrinos and neutrinos produced by the explosion of supernovae. The core of the
experiment is made of 20 kton Liquid Scintillator whose scintillation light is seen by almost 20000 large size (20") photo-multipliers
with high quantum efficiency, and by 25000 small size (3") photo-multipliers. The raw data will have to be further processed to
reconstruct the proper observables and for this task deep neural networks will be adopted for neutrino energy reconstruction. The
techniques are very powerful and allow to discriminate in an efficient way signal from background.ope
JUNO's Perspective for Geoneutrinos
No full textJiangmen Underground Neutrino Observatory (JUNO) is a neutrino experiment being built in Southern China to measure neutrinos produced in nuclear power plants at a distance of 52.5 km. Having the main goal to improve the knowledge about neutrino oscillations, fundamental properties of these particles, JUNO will also be able to observe neutrinos of natural origin such as from the Sun, supernovae, the Earth atmosphere and its interior. The latter are called geoneutrinos and can serve as a proxy for investigation of the Earth’s radiogenic heat. Using inverse beta-decay as the detection channel, JUNO is sensitive to geoneutrinos produced in beta-decays of U-238 and Th-232 radioactive isotopes. JUNO will collect in one year about 400 geoneutrinos, what is more than the present-day statistics measured by Borexino and KamLAND together. This talk will report the JUNO expected sensitivity to geoneutrino measurement including some preliminary results of an updated analysis
Preferential attachment with choice based edge-step
No full textWe study the asymptotic behavior of the maximum degree in the preferential
attachment model with a choice-based edge-step. We add vertex type to the model
and prove, among others types of behavior, the effect of condensation on
multiple vertices with different types
- …
