Using graph neural networks to reconstruct charged pion showers in the CMS High Granularity Calorimeter

The CMS HGCAL collaboration; The CALICE AHCAL collaborations

doi:10.1088/1748-0221/19/11/P11025

Using graph neural networks to reconstruct charged pion showers in the CMS High Granularity Calorimeter

The CMS HGCAL collaboration
, The CALICE AHCAL collaborations

University of Maryland, College Park
Georgian Technical University
Florida State University
National Central University
European Organization for Nuclear Research
Tata Institute of Fundamental Research, Mumbai
Indian Institute of Science Education and Research (IISER)
Quaid-i-Azam University
Texas Tech University
Yildiz Technical University
Bogazici University
Istanbul University
c/o DESY
Carnegie Mellon University
King Abdullah University of Science and Technology
Fermi National Accelerator Laboratory
Institute of Meteorology and Climate Research
Northern Illinois University
Lebanese University
Cukurova University
Imperial College London
University of Alabama
University of Wisconsin-Madison
Ip Paris
Indian Institute of Technology Madras
Institut für Hochenergiephysik
Universite Paris-Saclay
University of Rochester
Saha Institute of Nuclear Physics
University of California, Santa Barbara
University of Montenegro
Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of Split
Istanbul Technical University
National Taiwan University
National Technical University of Athens
University of Bristol
University of Minnesota Twin Cities
Nanjing Normal University
Massachusetts Institute of Technology
Academy of Science of Ukraine
University of Iowa
Baylor University
Institute of High Energy Physics, Chinese Academy of Sciences
University Malaya
Women and Infants Hospital of Rhode Island-Warren Alpert Medical School of Brown University
University of Helsinki
Northwestern University
University of Notre Dame
California Institute of Technology
University of Dundee
Kansas State University
Université Libre de Bruxelles
Riga Technical University
Kharkov Institute of Physics and Technology
Tsinghua University
Zhejiang University
Universität Hamburg
Bethel University
University of Milano-Bicocca
LIP - Lisboa
RWTH Aachen University
Boston University
Faculty of Science, University of Split
University of Bahrain

Résultats de recherche: Contribution à un journal › Article › Revue par des pairs

Résumé

A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadronic section. The shower reconstruction method is based on graph neural networks and it makes use of a dynamic reduction network architecture. It is shown that the algorithm is able to capture and mitigate the main effects that normally hinder the reconstruction of hadronic showers using classical reconstruction methods, by compensating for fluctuations in the multiplicity, energy, and spatial distributions of the shower's constituents. The performance of the algorithm is evaluated using test beam data collected in 2018 prototype of the CMS HGCAL accompanied by a section of the CALICE AHCAL prototype. The capability of the method to mitigate the impact of energy leakage from the calorimeter is also demonstrated.

langue originale	Anglais
Numéro d'article	P11025
journal	Journal of Instrumentation
Volume	19
Numéro de publication	11
Les DOIs	https://doi.org/10.1088/1748-0221/19/11/P11025
état	Publié - 1 nov. 2024

Accès au document

10.1088/1748-0221/19/11/P11025

Contient cette citation

@article{3edd87c3eec547f089b6a9216c408b5d,

title = "Using graph neural networks to reconstruct charged pion showers in the CMS High Granularity Calorimeter",

abstract = "A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadronic section. The shower reconstruction method is based on graph neural networks and it makes use of a dynamic reduction network architecture. It is shown that the algorithm is able to capture and mitigate the main effects that normally hinder the reconstruction of hadronic showers using classical reconstruction methods, by compensating for fluctuations in the multiplicity, energy, and spatial distributions of the shower's constituents. The performance of the algorithm is evaluated using test beam data collected in 2018 prototype of the CMS HGCAL accompanied by a section of the CALICE AHCAL prototype. The capability of the method to mitigate the impact of energy leakage from the calorimeter is also demonstrated.",

keywords = "Calorimeters, Pattern recognition, cluster finding, calibration and fitting methods, Performance of High Energy Physics Detectors, Si microstrip and pad detectors",

author = "\{The CMS HGCAL collaboration\} and \{The CALICE AHCAL collaborations\} and M. Aamir and G. Adamov and T. Adams and C. Adloff and S. Afanasiev and C. Agrawal and C. Agrawal and A. Ahmad and H. Ahmed and S. Akbar and N. Akchurin and B. Akgul and B. Akgun and R. Akpinar and E. Aktas and \{Al Kadhim\}, A. and V. Alexakhin and J. Alimena and J. Alison and A. Alpana and W. Alshehri and \{Alvarez Dominguez\}, P. and M. Alyari and C. Amendola and R. Amir and S. Andersen and Y. Andreev and P. Antoszczuk and U. Aras and L. Ardila and P. Aspell and M. Avila and I. Awad and O. Aydilek and Z. Azimi and \{Aznar Pretel\}, A. and O. Bach and R. Bainbridge and A. Bakshi and B. Bam and S. Banerjee and D. Barney and O. Bayraktar and A. Cappati and O. Davignon and \{De Wit\}, A. and C. Ochando and R. Salerno and A. Zabi and A. Zghiche",

note = "Publisher Copyright: {\textcopyright} 2024 CERN.",

year = "2024",

month = nov,

day = "1",

doi = "10.1088/1748-0221/19/11/P11025",

language = "English",

volume = "19",

journal = "Journal of Instrumentation",

issn = "1748-0221",

number = "11",

}

TY - JOUR

T1 - Using graph neural networks to reconstruct charged pion showers in the CMS High Granularity Calorimeter

AU - The CMS HGCAL collaboration

AU - The CALICE AHCAL collaborations

AU - Aamir, M.

AU - Adamov, G.

AU - Adams, T.

AU - Adloff, C.

AU - Afanasiev, S.

AU - Agrawal, C.

AU - Ahmad, A.

AU - Ahmed, H.

AU - Akbar, S.

AU - Akchurin, N.

AU - Akgul, B.

AU - Akgun, B.

AU - Akpinar, R.

AU - Aktas, E.

AU - Al Kadhim, A.

AU - Alexakhin, V.

AU - Alimena, J.

AU - Alison, J.

AU - Alpana, A.

AU - Alshehri, W.

AU - Alvarez Dominguez, P.

AU - Alyari, M.

AU - Amendola, C.

AU - Amir, R.

AU - Andersen, S.

AU - Andreev, Y.

AU - Antoszczuk, P.

AU - Aras, U.

AU - Ardila, L.

AU - Aspell, P.

AU - Avila, M.

AU - Awad, I.

AU - Aydilek, O.

AU - Azimi, Z.

AU - Aznar Pretel, A.

AU - Bach, O.

AU - Bainbridge, R.

AU - Bakshi, A.

AU - Bam, B.

AU - Banerjee, S.

AU - Barney, D.

AU - Bayraktar, O.

AU - Cappati, A.

AU - Davignon, O.

AU - De Wit, A.

AU - Ochando, C.

AU - Salerno, R.

AU - Zabi, A.

AU - Zghiche, A.

PY - 2024/11/1

Y1 - 2024/11/1

N2 - A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadronic section. The shower reconstruction method is based on graph neural networks and it makes use of a dynamic reduction network architecture. It is shown that the algorithm is able to capture and mitigate the main effects that normally hinder the reconstruction of hadronic showers using classical reconstruction methods, by compensating for fluctuations in the multiplicity, energy, and spatial distributions of the shower's constituents. The performance of the algorithm is evaluated using test beam data collected in 2018 prototype of the CMS HGCAL accompanied by a section of the CALICE AHCAL prototype. The capability of the method to mitigate the impact of energy leakage from the calorimeter is also demonstrated.

AB - A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadronic section. The shower reconstruction method is based on graph neural networks and it makes use of a dynamic reduction network architecture. It is shown that the algorithm is able to capture and mitigate the main effects that normally hinder the reconstruction of hadronic showers using classical reconstruction methods, by compensating for fluctuations in the multiplicity, energy, and spatial distributions of the shower's constituents. The performance of the algorithm is evaluated using test beam data collected in 2018 prototype of the CMS HGCAL accompanied by a section of the CALICE AHCAL prototype. The capability of the method to mitigate the impact of energy leakage from the calorimeter is also demonstrated.

KW - Calorimeters

KW - Pattern recognition, cluster finding, calibration and fitting methods

KW - Performance of High Energy Physics Detectors

KW - Si microstrip and pad detectors

U2 - 10.1088/1748-0221/19/11/P11025

DO - 10.1088/1748-0221/19/11/P11025

M3 - Article

AN - SCOPUS:85211203778

SN - 1748-0221

VL - 19

JO - Journal of Instrumentation

JF - Journal of Instrumentation

IS - 11

M1 - P11025

ER -

Using graph neural networks to reconstruct charged pion showers in the CMS High Granularity Calorimeter

Résumé

Accès au document

Empreinte digitale

Contient cette citation