Cargando…

Progress towards an improved particle flow algorithm at CMS with machine learning

The particle-flow (PF) algorithm, which infers particles based on tracks and calorimeter clusters, is of central importance to event reconstruction in the CMS experiment at the CERN LHC, and has been a focus of development in light of planned Phase-2 running conditions with an increased pileup and d...

Descripción completa

Detalles Bibliográficos
Autores principales: Mokhtar, Farouk, Pata, Joosep, Duarte, Javier, Wulff, Eric, Pierini, Maurizio, Vlimant, Jean-Roch
Lenguaje:eng
Publicado: 2023
Materias:
Acceso en línea:http://cds.cern.ch/record/2856311
_version_ 1780977498556727296
author Mokhtar, Farouk
Pata, Joosep
Duarte, Javier
Wulff, Eric
Pierini, Maurizio
Vlimant, Jean-Roch
author_facet Mokhtar, Farouk
Pata, Joosep
Duarte, Javier
Wulff, Eric
Pierini, Maurizio
Vlimant, Jean-Roch
author_sort Mokhtar, Farouk
collection CERN
description The particle-flow (PF) algorithm, which infers particles based on tracks and calorimeter clusters, is of central importance to event reconstruction in the CMS experiment at the CERN LHC, and has been a focus of development in light of planned Phase-2 running conditions with an increased pileup and detector granularity. In recent years, the machine learned particle-flow (MLPF) algorithm, a graph neural network that performs PF reconstruction, has been explored in CMS, with the possible advantages of directly optimizing for the physical quantities of interest, being highly reconfigurable to new conditions, and being a natural fit for deployment to heterogeneous accelerators. We discuss progress in CMS towards an improved implementation of the MLPF reconstruction, now optimized using generator/simulation-level particle information as the target for the first time. This paves the way to potentially improving the detector response in terms of physical quantities of interest. We describe the simulation-based training target, progress and studies on event-based loss terms, details on the model hyperparameter tuning, as well as physics validation with respect to the current PF algorithm in terms of high-level physical quantities such as the jet and missing transverse momentum resolutions. We find that the MLPF algorithm, trained on a generator/simulator level particle information for the first time, results in broadly compatible particle and jet reconstruction performance with the baseline PF, setting the stage for improving the physics performance by additional training statistics and model tuning.
id cern-2856311
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2023
record_format invenio
spelling cern-28563112023-04-21T09:39:15Zhttp://cds.cern.ch/record/2856311engMokhtar, FaroukPata, JoosepDuarte, JavierWulff, EricPierini, MaurizioVlimant, Jean-RochProgress towards an improved particle flow algorithm at CMS with machine learningphysics.ins-detDetectors and Experimental Techniqueshep-exParticle Physics - Experimentcs.LGComputing and Computersphysics.data-anOther Fields of PhysicsThe particle-flow (PF) algorithm, which infers particles based on tracks and calorimeter clusters, is of central importance to event reconstruction in the CMS experiment at the CERN LHC, and has been a focus of development in light of planned Phase-2 running conditions with an increased pileup and detector granularity. In recent years, the machine learned particle-flow (MLPF) algorithm, a graph neural network that performs PF reconstruction, has been explored in CMS, with the possible advantages of directly optimizing for the physical quantities of interest, being highly reconfigurable to new conditions, and being a natural fit for deployment to heterogeneous accelerators. We discuss progress in CMS towards an improved implementation of the MLPF reconstruction, now optimized using generator/simulation-level particle information as the target for the first time. This paves the way to potentially improving the detector response in terms of physical quantities of interest. We describe the simulation-based training target, progress and studies on event-based loss terms, details on the model hyperparameter tuning, as well as physics validation with respect to the current PF algorithm in terms of high-level physical quantities such as the jet and missing transverse momentum resolutions. We find that the MLPF algorithm, trained on a generator/simulator level particle information for the first time, results in broadly compatible particle and jet reconstruction performance with the baseline PF, setting the stage for improving the physics performance by additional training statistics and model tuning.arXiv:2303.17657oai:cds.cern.ch:28563112023-03-30
spellingShingle physics.ins-det
Detectors and Experimental Techniques
hep-ex
Particle Physics - Experiment
cs.LG
Computing and Computers
physics.data-an
Other Fields of Physics
Mokhtar, Farouk
Pata, Joosep
Duarte, Javier
Wulff, Eric
Pierini, Maurizio
Vlimant, Jean-Roch
Progress towards an improved particle flow algorithm at CMS with machine learning
title Progress towards an improved particle flow algorithm at CMS with machine learning
title_full Progress towards an improved particle flow algorithm at CMS with machine learning
title_fullStr Progress towards an improved particle flow algorithm at CMS with machine learning
title_full_unstemmed Progress towards an improved particle flow algorithm at CMS with machine learning
title_short Progress towards an improved particle flow algorithm at CMS with machine learning
title_sort progress towards an improved particle flow algorithm at cms with machine learning
topic physics.ins-det
Detectors and Experimental Techniques
hep-ex
Particle Physics - Experiment
cs.LG
Computing and Computers
physics.data-an
Other Fields of Physics
url http://cds.cern.ch/record/2856311
work_keys_str_mv AT mokhtarfarouk progresstowardsanimprovedparticleflowalgorithmatcmswithmachinelearning
AT patajoosep progresstowardsanimprovedparticleflowalgorithmatcmswithmachinelearning
AT duartejavier progresstowardsanimprovedparticleflowalgorithmatcmswithmachinelearning
AT wulfferic progresstowardsanimprovedparticleflowalgorithmatcmswithmachinelearning
AT pierinimaurizio progresstowardsanimprovedparticleflowalgorithmatcmswithmachinelearning
AT vlimantjeanroch progresstowardsanimprovedparticleflowalgorithmatcmswithmachinelearning