Please use this identifier to cite or link to this item: http://elea.unisa.it:8080/xmlui/handle/10556/3085
Title: Event triggering and deep learning for particle identification in KM3NeT
Authors: De Sio, Chiara
Scarpa, Roberto
Bozza, Cristiano
Keywords: KM3Net
Deep learning
Issue Date: 11-Apr-2018
Publisher: Universita degli studi di Salerno
Abstract: Neutrino astronomy experiments like KM3NeT allow to survey the Universe leveraging the properties of neutrinos of being electrically neutral and weakly interacting particles, making them a suitable messenger. Observing neutrino emission in association with electromagnetic radiation allows evaluating models for the acceleration of particles occurring in high energy sources such as Supernovae or Active Galactic Nuclei. This is the main goal of the ARCA project in KM3NeT. In addition, KM3NeT has a program for lower energy neutrinos called ORCA, aimed at distinguishing between the scenarios of “normal hierarchy” and “inverted hierarchy” for neutrino mass eigenstates. The KM3NeT Collaboration is currently building a network of three Cherenkov telescopes in the Mediterranean sea, in deep water off the coasts of Capopassero, Italy; Toulon, France, and Pylos, Greece. The water overburden shields the detectors from down-going charged particles produced by the interactions of cosmic rays in the atmosphere, while up-going neutrinos that cross the Earth are the target of the observation. Cosmic rays are a background to the KM3NeT signal, usually discarded by directional information. Nevertheless, they provide a reliable reference to calibrate the detector and work out its effective operating parameters, namely direction and energy of the incoming particles. Estimation of tracking capabilities is directly connected to the evaluation of the ability of the experiment to detect astrophysical point-like sources, i.e. its discovery potential. Being able to distinguish among the three neutrino flavours, or between neutrinos and muons, as well as estimating the neutrino direction and energy, are the main goals of such experiments. Trigger and reconstruction algorithms are designed to separate the signal from background and to provide an estimation for the above mentioned quantities, respectively... [edited by author]
Description: 2016 - 2017
URI: http://hdl.handle.net/10556/3085
http://dx.doi.org/10.14273/unisa-1369
Appears in Collections:Matematica, Fisica ed Applicazioni

Files in This Item:
File Description SizeFormat 
tesi C. De Sio.pdftesi di dottorato16,14 MBAdobe PDFView/Open
abstract in inglese C. De Sio.pdfabstract in inglese a cura dell’autore371,78 kBAdobe PDFView/Open
abstract in italiano C. De Sio.pdfabstract in italiano a cura dell’autore276,93 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.