PreprintCC BY 4.0HAL
Calibration of electromagnetic shower features in the CMS calorimeter with machine-learning techniques
Abstract
Monte Carlo simulations are used extensively in high-energy particle physics analyses. However, an incomplete description of the underlying physics can lead to significant discrepancies between observables measured in simulated and collision data. To mitigate potential biases arising from such mismodelling, it is essential to calibrate simulations to data. This paper presents two novel calibration methods based on machine-learning techniques: a reweighting approach, which employs a classifier to learn the ratio of probability density functions between simulation and data; and a normalising-flow approach, which learns a high-dimensional transformation to map simulation to data. Compared to traditional calibration methods, both approaches offer continuous, unbinned corrections across high-dimensional feature spaces, enabling improved global agreement with data. The techniques are demonstrated in the context of correcting the features of simulated electromagnetic showers in the CMS calorimeters, using proton-proton collision data collected during 2022 at √s = 13.6 TeV, corresponding to an integrated luminosity of 26.7 fb−1. The strengths and limitations of the two methods are compared.
Related
- Application of sequence learning for predicting radiation damage of the CMS electromagnetic calorimeter
- Enhancing electromagnetic calorimeter signal reconstruction with machine learning-based noise discrimination
- Kinematic Fitting of electromagnetic calorimeter data - an improved method
- Photon Calibration Techniques for High Resolution Cryogenic Detectors
- Understanding Energy Dependent Hadronic Calorimeter Response from a Machine Learning Perspective
BibTeX
@misc{gevorgyan2026calibration,
title = {{Calibration of electromagnetic shower features in the CMS calorimeter with machine-learning techniques}},
author = {Arzunik Gevorgyan and Aram Hayrapetyan and Vladimir Makarenko and Armen Tumasyan and Priya Sajid Hussain and Mangesh Sonawane and Wolfgang Adam and Lisa Benato and Thomas Bergauer and Marko Dragicevic and Stefan Gundacker and Ali Kaan Guven and Manfred Jeitler and Natascha Krammer and Ang Li and Dietrich Liko and Mark Nicholas Matthewman and Andreas Pfeiffer and Jochen Schieck and Robert Schoefbeck and Maryam Shooshtari and Niels van den Bossche and Wolfgang Waltenberger and Claudia Wulz and Tahys Janssen and Daniel Ocampo Henao and Thomas van Laer and Pierre van Mechelen and Donya Ahmadi and Jas Bierkens and Nordin Breugelmans and Soumya Dansana and Alexandre de Moor and Martin Delcourt and Stef Duponcheel and Chirayu Gupta and Felix Johannes Heyen and Yanwen Hong and Kai Kang and Pavlo Kashko and Steven Lowette and Inna Makarenko and Saranya Nandakumar and Jeremi Niedziela and Stefaan Tavernier and Michael Tytgat and Gerrit van Onsem and Senne van Putte and Tiepolo Wybouw and Aya Beshr and Bugra Bilin and Franco Caviglia Roman and Barbara Clerbaux and Aloke Kumar Das and Isabelle de Bruyn and Gilles de Lentdecker and Eliott Ducarme and Hugues Evard and Laurent Favart and Ilia Kalaitzidou},
howpublished = {HAL},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.15377},
eprint = {2609.15377},
archivePrefix = {arXiv},
url = {https://hal.science/hal-05763105v1},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Calibration of electromagnetic shower features in the CMS calorimeter with machine-learning techniques
AU - Arzunik Gevorgyan
AU - Aram Hayrapetyan
AU - Vladimir Makarenko
AU - Armen Tumasyan
AU - Priya Sajid Hussain
AU - Mangesh Sonawane
AU - Wolfgang Adam
AU - Lisa Benato
AU - Thomas Bergauer
AU - Marko Dragicevic
AU - Stefan Gundacker
AU - Ali Kaan Guven
AU - Manfred Jeitler
AU - Natascha Krammer
AU - Ang Li
AU - Dietrich Liko
AU - Mark Nicholas Matthewman
AU - Andreas Pfeiffer
AU - Jochen Schieck
AU - Robert Schoefbeck
AU - Maryam Shooshtari
AU - Niels van den Bossche
AU - Wolfgang Waltenberger
AU - Claudia Wulz
AU - Tahys Janssen
AU - Daniel Ocampo Henao
AU - Thomas van Laer
AU - Pierre van Mechelen
AU - Donya Ahmadi
AU - Jas Bierkens
AU - Nordin Breugelmans
AU - Soumya Dansana
AU - Alexandre de Moor
AU - Martin Delcourt
AU - Stef Duponcheel
AU - Chirayu Gupta
AU - Felix Johannes Heyen
AU - Yanwen Hong
AU - Kai Kang
AU - Pavlo Kashko
AU - Steven Lowette
AU - Inna Makarenko
AU - Saranya Nandakumar
AU - Jeremi Niedziela
AU - Stefaan Tavernier
AU - Michael Tytgat
AU - Gerrit van Onsem
AU - Senne van Putte
AU - Tiepolo Wybouw
AU - Aya Beshr
AU - Bugra Bilin
AU - Franco Caviglia Roman
AU - Barbara Clerbaux
AU - Aloke Kumar Das
AU - Isabelle de Bruyn
AU - Gilles de Lentdecker
AU - Eliott Ducarme
AU - Hugues Evard
AU - Laurent Favart
AU - Ilia Kalaitzidou
PY - 2026
DA - 2026/09/14
PB - HAL
DO - 10.48550/arxiv.2609.15377
UR - https://hal.science/hal-05763105v1
AB - Monte Carlo simulations are used extensively in high-energy particle physics analyses. However, an incomplete description of the underlying physics can lead to significant discrepancies between observables measured in simulated and collision data. To mitigate potential biases arising from such mismodelling, it is essential to calibrate simulations to data. This paper presents two novel calibration methods based on machine-learning techniques: a reweighting approach, which employs a classifier to learn the ratio of probability density functions between simulation and data; and a normalising-flow approach, which learns a high-dimensional transformation to map simulation to data. Compared to traditional calibration methods, both approaches offer continuous, unbinned corrections across high-dimensional feature spaces, enabling improved global agreement with data. The techniques are demonstrated in the context of correcting the features of simulated electromagnetic showers in the CMS calorimeters, using proton-proton collision data collected during 2022 at √s = 13.6 TeV, corresponding to an integrated luminosity of 26.7 fb^−1. The strengths and limitations of the two methods are compared.
ER -