Deep Neural Networks for combined neutrino energy estimate with KM3NeT/ORCA6

KM3NeT Collaboration

Deep Neural Networks for combined neutrino energy estimate with KM3NeT/ORCA6

KM3NeT Collaboration

Physics

Université de Paris
National Institute for Nuclear Physics
Université de Strasbourg
Université de Haute-Alsace
IFIC (CSIC-Universitat de València)
Aix-Marseille Université
University of Naples Federico II
Polytechnic University of Catalonia
Demokritos National Centre for Scientific Research
University of Granada
Nantes Université
Polytechnic University of Valencia
Mohammed V University in Rabat
Université de Caen
Czech Technical University in Prague
Comenius University
National Institute for Subatomic Physics
University of Bologna
University of Campania Luigi Vanvitelli
University of Hull
North West University
Mohamed I University
University of Salerno
Institute for Space Sciences
University of Amsterdam
Netherlands Organisation for Applied Scientific Research
University of Rome La Sapienza
Cadi Ayyad University
Friedrich-Alexander University Erlangen-Nürnberg
University of the Witwatersrand
University of Catania
International Centre for Radio Astronomy Research
University of Würzburg
Western Sydney University
LPC
University of Genoa
Royal Netherlands Institute for Sea Research - NIOZ
National Centre for Nuclear Research
Nicolaus Copernicus Astronomical Center of the Polish Academy of Sciences
Ivane Javakhishvili Tbilisi State University
Institut universitaire de France
The University of Georgia, Tbilisi
IN2P3 - Institut National de Physique Nucléaire et de Physique Des Particules
Leiden University
Université Montpellier 2

Research output: Contribution to journal › Conference article › peer-review

Abstract

KM3NeT/ORCA is a large-volume water-Cherenkov neutrino detector, currently under construction at the bottom of the Mediterranean Sea at a depth of 2450 meters. The main research goal of ORCA is the measurement of the neutrino mass ordering and the atmospheric neutrino oscillation parameters. Additionally, the detector is also sensitive to a wide variety of phenomena including non-standard neutrino interactions, sterile neutrinos, and neutrino decay. This contribution describes the use of a machine learning framework for building Deep Neural Networks (DNN) which combine multiple energy estimates to generate a more precise reconstructed neutrino energy. The model is optimized to improve the oscillation analysis based on a data sample of 433 kton-years of KM3NeT/ORCA with 6 detection units. The performance of the model is evaluated by determining the sensitivity to oscillation parameters in comparison with the standard energy reconstruction method of maximizing a likelihood function. The results show that the DNN is able to provide a better energy estimate with lower bias in the context of oscillation analyses.

Original language	English
Article number	1035
Journal	Proceedings of Science
Volume	444
Publication status	Published - 27 Sept 2024
Event	38th International Cosmic Ray Conference, ICRC 2023 - Nagoya, Japan Duration: 26 Jul 2023 → 3 Aug 2023

ASJC Scopus subject areas

Multidisciplinary

Cite this

@article{df1b0cb23e654b0bb021cb9e80dc2456,

title = "Deep Neural Networks for combined neutrino energy estimate with KM3NeT/ORCA6",

abstract = "KM3NeT/ORCA is a large-volume water-Cherenkov neutrino detector, currently under construction at the bottom of the Mediterranean Sea at a depth of 2450 meters. The main research goal of ORCA is the measurement of the neutrino mass ordering and the atmospheric neutrino oscillation parameters. Additionally, the detector is also sensitive to a wide variety of phenomena including non-standard neutrino interactions, sterile neutrinos, and neutrino decay. This contribution describes the use of a machine learning framework for building Deep Neural Networks (DNN) which combine multiple energy estimates to generate a more precise reconstructed neutrino energy. The model is optimized to improve the oscillation analysis based on a data sample of 433 kton-years of KM3NeT/ORCA with 6 detection units. The performance of the model is evaluated by determining the sensitivity to oscillation parameters in comparison with the standard energy reconstruction method of maximizing a likelihood function. The results show that the DNN is able to provide a better energy estimate with lower bias in the context of oscillation analyses.",

author = "\{KM3NeT Collaboration\} and Mart{\'i}nez, \{Santiago Pe{\~n}a\} and S. Aiello and A. Albert and \{Alves Garre\}, S. and Z. Aly and A. Ambrosone and F. Ameli and M. Andre and E. Androutsou and M. Anguita and L. Aphecetche and M. Ardid and S. Ardid and H. Atmani and J. Aublin and L. Bailly-Salins and Z. Barda{\v c}ov{\'a} and B. Baret and A. Bariego-Quintana and \{Basegmez du Pree\}, S. and Y. Becherini and M. Bendahman and F. Benfenati and M. Benhassi and Benoit, \{D. M.\} and E. Berbee and V. Bertin and S. Biagi and M. Boettcher and D. Bonanno and J. Boumaaza and M. Bouta and M. Bouwhuis and C. Bozza and Bozza, \{R. M.\} and H. Br{\^a}nza{\c s} and F. Bretaudeau and R. Bruijn and J. Brunner and R. Bruno and E. Buis and R. Buompane and J. Busto and B. Caiffi and D. Calvo and S. Campion and A. Capone and F. Carenini and V. Carretero and S. Razzaque",

note = "Publisher Copyright: {\textcopyright} Copyright owned by the author(s) under the terms of the Creative Commons.; 38th International Cosmic Ray Conference, ICRC 2023 ; Conference date: 26-07-2023 Through 03-08-2023",

year = "2024",

month = sep,

day = "27",

language = "English",

volume = "444",

journal = "Proceedings of Science",

issn = "1824-8039",

publisher = "Sissa Medialab Srl",

}

TY - JOUR

T1 - Deep Neural Networks for combined neutrino energy estimate with KM3NeT/ORCA6

AU - KM3NeT Collaboration

AU - Martínez, Santiago Peña

AU - Aiello, S.

AU - Albert, A.

AU - Alves Garre, S.

AU - Aly, Z.

AU - Ambrosone, A.

AU - Ameli, F.

AU - Andre, M.

AU - Androutsou, E.

AU - Anguita, M.

AU - Aphecetche, L.

AU - Ardid, M.

AU - Ardid, S.

AU - Atmani, H.

AU - Aublin, J.

AU - Bailly-Salins, L.

AU - Bardačová, Z.

AU - Baret, B.

AU - Bariego-Quintana, A.

AU - Basegmez du Pree, S.

AU - Becherini, Y.

AU - Bendahman, M.

AU - Benfenati, F.

AU - Benhassi, M.

AU - Benoit, D. M.

AU - Berbee, E.

AU - Bertin, V.

AU - Biagi, S.

AU - Boettcher, M.

AU - Bonanno, D.

AU - Boumaaza, J.

AU - Bouta, M.

AU - Bouwhuis, M.

AU - Bozza, C.

AU - Bozza, R. M.

AU - Brânzaş, H.

AU - Bretaudeau, F.

AU - Bruijn, R.

AU - Brunner, J.

AU - Bruno, R.

AU - Buis, E.

AU - Buompane, R.

AU - Busto, J.

AU - Caiffi, B.

AU - Calvo, D.

AU - Campion, S.

AU - Capone, A.

AU - Carenini, F.

AU - Carretero, V.

AU - Razzaque, S.

PY - 2024/9/27

Y1 - 2024/9/27

N2 - KM3NeT/ORCA is a large-volume water-Cherenkov neutrino detector, currently under construction at the bottom of the Mediterranean Sea at a depth of 2450 meters. The main research goal of ORCA is the measurement of the neutrino mass ordering and the atmospheric neutrino oscillation parameters. Additionally, the detector is also sensitive to a wide variety of phenomena including non-standard neutrino interactions, sterile neutrinos, and neutrino decay. This contribution describes the use of a machine learning framework for building Deep Neural Networks (DNN) which combine multiple energy estimates to generate a more precise reconstructed neutrino energy. The model is optimized to improve the oscillation analysis based on a data sample of 433 kton-years of KM3NeT/ORCA with 6 detection units. The performance of the model is evaluated by determining the sensitivity to oscillation parameters in comparison with the standard energy reconstruction method of maximizing a likelihood function. The results show that the DNN is able to provide a better energy estimate with lower bias in the context of oscillation analyses.

AB - KM3NeT/ORCA is a large-volume water-Cherenkov neutrino detector, currently under construction at the bottom of the Mediterranean Sea at a depth of 2450 meters. The main research goal of ORCA is the measurement of the neutrino mass ordering and the atmospheric neutrino oscillation parameters. Additionally, the detector is also sensitive to a wide variety of phenomena including non-standard neutrino interactions, sterile neutrinos, and neutrino decay. This contribution describes the use of a machine learning framework for building Deep Neural Networks (DNN) which combine multiple energy estimates to generate a more precise reconstructed neutrino energy. The model is optimized to improve the oscillation analysis based on a data sample of 433 kton-years of KM3NeT/ORCA with 6 detection units. The performance of the model is evaluated by determining the sensitivity to oscillation parameters in comparison with the standard energy reconstruction method of maximizing a likelihood function. The results show that the DNN is able to provide a better energy estimate with lower bias in the context of oscillation analyses.

UR - https://www.scopus.com/pages/publications/85212280633

M3 - Conference article

AN - SCOPUS:85212280633

SN - 1824-8039

VL - 444

JO - Proceedings of Science

JF - Proceedings of Science

M1 - 1035

T2 - 38th International Cosmic Ray Conference, ICRC 2023

Y2 - 26 July 2023 through 3 August 2023

ER -

Deep Neural Networks for combined neutrino energy estimate with KM3NeT/ORCA6

Abstract

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this