Particle identification at MeV energies in JUNO

JUNO is a multi-purpose neutrino experiment currently under construction in Jiangmen, China. It primary aiming to determine the mass ordering. Moreover, its 20\,kt target makes it an ideal detector study neutrinos from various sources, including nuclear reactors, Earth and atmosphere, Sun, even supernovae. Due small cross section of interactions, event rate experiments limited. In order maximize signal-to-noise ratio, extremely important control background levels. this paper we discuss potential particle identification JUNO, underlying principles possible areas application experiment. While presented concepts can be transferred any large liquid scintillator detector, our methods are evaluated specifically for results mainly driven by high optical photon yield 1,200 photo electrons per MeV deposited energy. investigate discrimination, several pairings analysed, i.e. $\alpha/\beta$, $p/\beta$, $e^+/e^-$, $e^-/\gamma$. We compare discrimination performance advanced analytical techniques based on neural networks topological reconstruction keeping standard Gatti filter as reference. use Monte Carlo samples generated physically motivated energy intervals. dependence cuts energy, radial position, PMT time resolution, dark noise. The show excellent $\alpha/\beta$ $p/\beta$ with method network. Furthermore, $e^+/e^-$ $e^-/\gamma$ partly distinguished means network statistical basis. Especially latter case, proved very successful.