Event reconstruction for KM3NeT/ORCA using convolutional neural networks

S. Aiello; A. Albert; S. Alves Garre; Z. Aly; F. Ameli; M. Andre; G. Androulakis; M. Anghinolfi; M. Anguita; G. Anton; M. Ardid; J. Aublin; C. Bagatelas; G. Barbarino; B. Baret; S. Basegmez du Pree; M. Bendahman; E. Berbee; A. M. van den Berg; V. Bertin; S. Biagi; A. Biagioni; M. Bissinger; M. Boettcher; J. Boumaaza; M. Bouta; M. Bouwhuis; C. Bozza; H. BrÃ¢nzaş; R. Bruijn; J. Brunner; E. Buis; R. Buompane; J. Busto; B. Caiffi; D. Calvo; A. Capone; V. Carretero; P. Castaldi; S. Celli; M. Chabab; N. Chau; A. Chen; S. Cherubini; V. Chiarella; T. Chiarusi; M. Circella; R. Cocimano; J. A. B. Coelho; A. Coleiro; M. Colomer Molla; R. Coniglione; P. Coyle; A. Creusot; G. Cuttone; A. D'Onofrio; R. Dallier; M. De Palma; I. Di Palma; A. F. Diaz; D. Diego-Tortosa; C. Distefano; A. Domi; R. Donü; C. Donzaud; D. Dornic; M. Dörr; D. Drouhin; T. Eberl; A. Eddyamoui; T. van Eden; D. van Eijk; I. El Bojaddaini; D. Elsaesser; A. Enzenhöfer; V. Espinosa Roselló; P. Fermani; G. Ferrara; M. D. Filipović; null et al

doi:10.1088/1748-0221/15/10/P10005

Event reconstruction for KM3NeT/ORCA using convolutional neural networks

S. Aiello, A. Albert, S. Alves Garre, Z. Aly, F. Ameli, M. Andre, G. Androulakis, M. Anghinolfi, M. Anguita, G. Anton, M. Ardid, J. Aublin, C. Bagatelas, G. Barbarino, B. Baret, S. Basegmez du Pree, M. Bendahman, E. Berbee, A. M. van den Berg, V. BertinS. Biagi, A. Biagioni, M. Bissinger, M. Boettcher, J. Boumaaza, M. Bouta, M. Bouwhuis, C. Bozza, H. BrÃ¢nzaş, R. Bruijn, J. Brunner, E. Buis, R. Buompane, J. Busto, B. Caiffi, D. Calvo, A. Capone, V. Carretero, P. Castaldi, S. Celli, M. Chabab, N. Chau, A. Chen, S. Cherubini, V. Chiarella, T. Chiarusi, M. Circella, R. Cocimano, J. A. B. Coelho, A. Coleiro, M. Colomer Molla, R. Coniglione, P. Coyle, A. Creusot, G. Cuttone, A. D'Onofrio, R. Dallier, M. De Palma, I. Di Palma, A. F. Diaz, D. Diego-Tortosa, C. Distefano, A. Domi, R. Donü, C. Donzaud, D. Dornic, M. Dörr, D. Drouhin, T. Eberl, A. Eddyamoui, T. van Eden, D. van Eijk, I. El Bojaddaini, D. Elsaesser, A. Enzenhöfer, V. Espinosa Roselló, P. Fermani, G. Ferrara, M. D. Filipović, et al

Western Sydney University

Research output: Contribution to journal › Article › peer-review

22 Citations (Scopus)

Abstract

The KM3NeT research infrastructure is currently under construction at two locations in the Mediterranean Sea. The KM3NeT/ORCA water-Cherenkov neutrino detector off the French coast will instrument several megatons of seawater with photosensors. Its main objective is the determination of the neutrino mass ordering. This work aims at demonstrating the general applicability of deep convolutional neural networks to neutrino telescopes, using simulated datasets for the KM3NeT/ORCA detector as an example. To this end, the networks are employed to achieve reconstruction and classification tasks that constitute an alternative to the analysis pipeline presented for KM3NeT/ORCA in the KM3NeT Letter of Intent. They are used to infer event reconstruction estimates for the energy, the direction, and the interaction point of incident neutrinos. The spatial distribution of Cherenkov light generated by charged particles induced in neutrino interactions is classified as shower- or track-like, and the main background processes associated with the detection of atmospheric neutrinos are recognized. Performance comparisons to machine-learning classification and maximum-likelihood reconstruction algorithms previously developed for KM3NeT/ORCA are provided. It is shown that this application of deep convolutional neural networks to simulated datasets for a large-volume neutrino telescope yields competitive reconstruction results and performance improvements with respect to classical approaches.

Original language	English
Article number	P10005
Number of pages	39
Journal	Journal of Instrumentation
Volume	15
Issue number	10
DOIs	https://doi.org/10.1088/1748-0221/15/10/P10005
Publication status	Published - Oct 2020

Bibliographical note

Publisher Copyright:
© 2020 The Author(s). Published by IOP Publishing Ltd on behalf of Sissa Medialab. Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Open Access - Access Right Statement

© 2020 The Author(s). Published by IOP Publishing Ltd on behalf of Sissa Medialab. Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence (https://creativecommons.org/licenses/by/4.0/). Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Keywords

Cherenkov counters
convolutions (mathematics)
detectors
machine learning
neural networks (computer science)
performance

Access to Document

10.1088/1748-0221/15/10/P10005

Cite this

Aiello, S., Albert, A., Garre, S. A., Aly, Z., Ameli, F., Andre, M., Androulakis, G., Anghinolfi, M., Anguita, M., Anton, G., Ardid, M., Aublin, J., Bagatelas, C., Barbarino, G., Baret, B., du Pree, S. B., Bendahman, M., Berbee, E., van den Berg, A. M., ... et al (2020). Event reconstruction for KM3NeT/ORCA using convolutional neural networks. Journal of Instrumentation, 15(10), Article P10005. https://doi.org/10.1088/1748-0221/15/10/P10005

@article{321da80eac4a4bed849b868171aee5cf,

title = "Event reconstruction for KM3NeT/ORCA using convolutional neural networks",

abstract = "The KM3NeT research infrastructure is currently under construction at two locations in the Mediterranean Sea. The KM3NeT/ORCA water-Cherenkov neutrino detector off the French coast will instrument several megatons of seawater with photosensors. Its main objective is the determination of the neutrino mass ordering. This work aims at demonstrating the general applicability of deep convolutional neural networks to neutrino telescopes, using simulated datasets for the KM3NeT/ORCA detector as an example. To this end, the networks are employed to achieve reconstruction and classification tasks that constitute an alternative to the analysis pipeline presented for KM3NeT/ORCA in the KM3NeT Letter of Intent. They are used to infer event reconstruction estimates for the energy, the direction, and the interaction point of incident neutrinos. The spatial distribution of Cherenkov light generated by charged particles induced in neutrino interactions is classified as shower- or track-like, and the main background processes associated with the detection of atmospheric neutrinos are recognized. Performance comparisons to machine-learning classification and maximum-likelihood reconstruction algorithms previously developed for KM3NeT/ORCA are provided. It is shown that this application of deep convolutional neural networks to simulated datasets for a large-volume neutrino telescope yields competitive reconstruction results and performance improvements with respect to classical approaches.",

keywords = "Cherenkov counters, convolutions (mathematics), detectors, machine learning, neural networks (computer science), performance",

author = "S. Aiello and A. Albert and Garre, {S. Alves} and Z. Aly and F. Ameli and M. Andre and G. Androulakis and M. Anghinolfi and M. Anguita and G. Anton and M. Ardid and J. Aublin and C. Bagatelas and G. Barbarino and B. Baret and {du Pree}, {S. Basegmez} and M. Bendahman and E. Berbee and {van den Berg}, {A. M.} and V. Bertin and S. Biagi and A. Biagioni and M. Bissinger and M. Boettcher and J. Boumaaza and M. Bouta and M. Bouwhuis and C. Bozza and H. Br{\~A}¢nza{\c s} and R. Bruijn and J. Brunner and E. Buis and R. Buompane and J. Busto and B. Caiffi and D. Calvo and A. Capone and V. Carretero and P. Castaldi and S. Celli and M. Chabab and N. Chau and A. Chen and S. Cherubini and V. Chiarella and T. Chiarusi and M. Circella and R. Cocimano and Coelho, {J. A. B.} and A. Coleiro and Molla, {M. Colomer} and R. Coniglione and P. Coyle and A. Creusot and G. Cuttone and A. D'Onofrio and R. Dallier and {De Palma}, M. and {Di Palma}, I. and Diaz, {A. F.} and D. Diego-Tortosa and C. Distefano and A. Domi and R. Don{\"u} and C. Donzaud and D. Dornic and M. D{\"o}rr and D. Drouhin and T. Eberl and A. Eddyamoui and {van Eden}, T. and {van Eijk}, D. and {El Bojaddaini}, I. and D. Elsaesser and A. Enzenh{\"o}fer and Rosell{\'o}, {V. Espinosa} and P. Fermani and G. Ferrara and Filipovi{\'c}, {M. D.} and {et al}",

note = "Publisher Copyright: {\textcopyright} 2020 The Author(s). Published by IOP Publishing Ltd on behalf of Sissa Medialab. Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.",

year = "2020",

month = oct,

doi = "10.1088/1748-0221/15/10/P10005",

language = "English",

volume = "15",

journal = "Journal of Instrumentation",

issn = "1748-0221",

publisher = "Institute of Physics",

number = "10",

}

Aiello, S, Albert, A, Garre, SA, Aly, Z, Ameli, F, Andre, M, Androulakis, G, Anghinolfi, M, Anguita, M, Anton, G, Ardid, M, Aublin, J, Bagatelas, C, Barbarino, G, Baret, B, du Pree, SB, Bendahman, M, Berbee, E, van den Berg, AM, Bertin, V, Biagi, S, Biagioni, A, Bissinger, M, Boettcher, M, Boumaaza, J, Bouta, M, Bouwhuis, M, Bozza, C, BrÃ¢nzaş, H, Bruijn, R, Brunner, J, Buis, E, Buompane, R, Busto, J, Caiffi, B, Calvo, D, Capone, A, Carretero, V, Castaldi, P, Celli, S, Chabab, M, Chau, N, Chen, A, Cherubini, S, Chiarella, V, Chiarusi, T, Circella, M, Cocimano, R, Coelho, JAB, Coleiro, A, Molla, MC, Coniglione, R, Coyle, P, Creusot, A, Cuttone, G, D'Onofrio, A, Dallier, R, De Palma, M, Di Palma, I, Diaz, AF, Diego-Tortosa, D, Distefano, C, Domi, A, Donü, R, Donzaud, C, Dornic, D, Dörr, M, Drouhin, D, Eberl, T, Eddyamoui, A, van Eden, T, van Eijk, D, El Bojaddaini, I, Elsaesser, D, Enzenhöfer, A, Roselló, VE, Fermani, P, Ferrara, G, Filipović, MD & et al 2020, 'Event reconstruction for KM3NeT/ORCA using convolutional neural networks', Journal of Instrumentation, vol. 15, no. 10, P10005. https://doi.org/10.1088/1748-0221/15/10/P10005

TY - JOUR

T1 - Event reconstruction for KM3NeT/ORCA using convolutional neural networks

AU - Aiello, S.

AU - Albert, A.

AU - Garre, S. Alves

AU - Aly, Z.

AU - Ameli, F.

AU - Andre, M.

AU - Androulakis, G.

AU - Anghinolfi, M.

AU - Anguita, M.

AU - Anton, G.

AU - Ardid, M.

AU - Aublin, J.

AU - Bagatelas, C.

AU - Barbarino, G.

AU - Baret, B.

AU - du Pree, S. Basegmez

AU - Bendahman, M.

AU - Berbee, E.

AU - van den Berg, A. M.

AU - Bertin, V.

AU - Biagi, S.

AU - Biagioni, A.

AU - Bissinger, M.

AU - Boettcher, M.

AU - Boumaaza, J.

AU - Bouta, M.

AU - Bouwhuis, M.

AU - Bozza, C.

AU - BrÃ¢nzaş, H.

AU - Bruijn, R.

AU - Brunner, J.

AU - Buis, E.

AU - Buompane, R.

AU - Busto, J.

AU - Caiffi, B.

AU - Calvo, D.

AU - Capone, A.

AU - Carretero, V.

AU - Castaldi, P.

AU - Celli, S.

AU - Chabab, M.

AU - Chau, N.

AU - Chen, A.

AU - Cherubini, S.

AU - Chiarella, V.

AU - Chiarusi, T.

AU - Circella, M.

AU - Cocimano, R.

AU - Coelho, J. A. B.

AU - Coleiro, A.

AU - Molla, M. Colomer

AU - Coniglione, R.

AU - Coyle, P.

AU - Creusot, A.

AU - Cuttone, G.

AU - D'Onofrio, A.

AU - Dallier, R.

AU - De Palma, M.

AU - Di Palma, I.

AU - Diaz, A. F.

AU - Diego-Tortosa, D.

AU - Distefano, C.

AU - Domi, A.

AU - Donü, R.

AU - Donzaud, C.

AU - Dornic, D.

AU - Dörr, M.

AU - Drouhin, D.

AU - Eberl, T.

AU - Eddyamoui, A.

AU - van Eden, T.

AU - van Eijk, D.

AU - El Bojaddaini, I.

AU - Elsaesser, D.

AU - Enzenhöfer, A.

AU - Roselló, V. Espinosa

AU - Fermani, P.

AU - Ferrara, G.

AU - Filipović, M. D.

AU - et al, null

N1 - Publisher Copyright: © 2020 The Author(s). Published by IOP Publishing Ltd on behalf of Sissa Medialab. Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

PY - 2020/10

Y1 - 2020/10

N2 - The KM3NeT research infrastructure is currently under construction at two locations in the Mediterranean Sea. The KM3NeT/ORCA water-Cherenkov neutrino detector off the French coast will instrument several megatons of seawater with photosensors. Its main objective is the determination of the neutrino mass ordering. This work aims at demonstrating the general applicability of deep convolutional neural networks to neutrino telescopes, using simulated datasets for the KM3NeT/ORCA detector as an example. To this end, the networks are employed to achieve reconstruction and classification tasks that constitute an alternative to the analysis pipeline presented for KM3NeT/ORCA in the KM3NeT Letter of Intent. They are used to infer event reconstruction estimates for the energy, the direction, and the interaction point of incident neutrinos. The spatial distribution of Cherenkov light generated by charged particles induced in neutrino interactions is classified as shower- or track-like, and the main background processes associated with the detection of atmospheric neutrinos are recognized. Performance comparisons to machine-learning classification and maximum-likelihood reconstruction algorithms previously developed for KM3NeT/ORCA are provided. It is shown that this application of deep convolutional neural networks to simulated datasets for a large-volume neutrino telescope yields competitive reconstruction results and performance improvements with respect to classical approaches.

AB - The KM3NeT research infrastructure is currently under construction at two locations in the Mediterranean Sea. The KM3NeT/ORCA water-Cherenkov neutrino detector off the French coast will instrument several megatons of seawater with photosensors. Its main objective is the determination of the neutrino mass ordering. This work aims at demonstrating the general applicability of deep convolutional neural networks to neutrino telescopes, using simulated datasets for the KM3NeT/ORCA detector as an example. To this end, the networks are employed to achieve reconstruction and classification tasks that constitute an alternative to the analysis pipeline presented for KM3NeT/ORCA in the KM3NeT Letter of Intent. They are used to infer event reconstruction estimates for the energy, the direction, and the interaction point of incident neutrinos. The spatial distribution of Cherenkov light generated by charged particles induced in neutrino interactions is classified as shower- or track-like, and the main background processes associated with the detection of atmospheric neutrinos are recognized. Performance comparisons to machine-learning classification and maximum-likelihood reconstruction algorithms previously developed for KM3NeT/ORCA are provided. It is shown that this application of deep convolutional neural networks to simulated datasets for a large-volume neutrino telescope yields competitive reconstruction results and performance improvements with respect to classical approaches.

KW - Cherenkov counters

KW - convolutions (mathematics)

KW - detectors

KW - machine learning

KW - neural networks (computer science)

KW - performance

UR - https://hdl.handle.net/1959.7/uws:58172

U2 - 10.1088/1748-0221/15/10/P10005

DO - 10.1088/1748-0221/15/10/P10005

M3 - Article

SN - 1748-0221

VL - 15

JO - Journal of Instrumentation

JF - Journal of Instrumentation

IS - 10

M1 - P10005

ER -

Event reconstruction for KM3NeT/ORCA using convolutional neural networks

Abstract

Bibliographical note

Open Access - Access Right Statement

Keywords

Access to Document

Other files and links

Fingerprint

Cite this