Learning to Remove Pileup at the LHC with Jet Images

We present the Pileup Mitgation with Machine Learning (PUMML) algorithm for pileup removal at the Large Hadron Collider (LHC) based on the jet images framework using state-of-the-art machine learning techniques. We demonstrate that our algorithm outperforms existing methods on a wide range of jet ob...

Descripción completa

Detalles Bibliográficos
Autores principales: Komiske, Patrick T, Metodiev, Eric M, Nachman, Benjamin, Schwartz, Matthew D
Lenguaje:eng
Publicado: 2018
Materias:
Computing and Computers
Particle Physics - Phenomenology
Acceso en línea:https://dx.doi.org/10.1088/1742-6596/1085/4/042010
http://cds.cern.ch/record/2669842
_version_ 1780962333214900224
author Komiske, Patrick T
Metodiev, Eric M
Nachman, Benjamin
Schwartz, Matthew D
author_facet Komiske, Patrick T
Metodiev, Eric M
Nachman, Benjamin
Schwartz, Matthew D
author_sort Komiske, Patrick T
collection CERN
description We present the Pileup Mitgation with Machine Learning (PUMML) algorithm for pileup removal at the Large Hadron Collider (LHC) based on the jet images framework using state-of-the-art machine learning techniques. We demonstrate that our algorithm outperforms existing methods on a wide range of jet observables up to pileup levels of 140 collisions per bunch crossing. We also investigate what aspects of the event our algorithms are utilizing by understanding the learned parameters of a simplified version of the model.
id oai-inspirehep.net-1699881
institution Organización Europea para la Investigación Nuclear
language eng
publishDate 2018
record_format invenio
spelling oai-inspirehep.net-16998812021-02-09T10:06:49Zdoi:10.1088/1742-6596/1085/4/042010http://cds.cern.ch/record/2669842engKomiske, Patrick TMetodiev, Eric MNachman, BenjaminSchwartz, Matthew DLearning to Remove Pileup at the LHC with Jet ImagesComputing and ComputersParticle Physics - PhenomenologyWe present the Pileup Mitgation with Machine Learning (PUMML) algorithm for pileup removal at the Large Hadron Collider (LHC) based on the jet images framework using state-of-the-art machine learning techniques. We demonstrate that our algorithm outperforms existing methods on a wide range of jet observables up to pileup levels of 140 collisions per bunch crossing. We also investigate what aspects of the event our algorithms are utilizing by understanding the learned parameters of a simplified version of the model.oai:inspirehep.net:16998812018
spellingShingle Computing and Computers
Particle Physics - Phenomenology
Komiske, Patrick T
Metodiev, Eric M
Nachman, Benjamin
Schwartz, Matthew D
Learning to Remove Pileup at the LHC with Jet Images
title Learning to Remove Pileup at the LHC with Jet Images
title_full Learning to Remove Pileup at the LHC with Jet Images
title_fullStr Learning to Remove Pileup at the LHC with Jet Images
title_full_unstemmed Learning to Remove Pileup at the LHC with Jet Images
title_short Learning to Remove Pileup at the LHC with Jet Images
title_sort learning to remove pileup at the lhc with jet images
topic Computing and Computers
Particle Physics - Phenomenology
url https://dx.doi.org/10.1088/1742-6596/1085/4/042010
http://cds.cern.ch/record/2669842
work_keys_str_mv AT komiskepatrickt learningtoremovepileupatthelhcwithjetimages
AT metodievericm learningtoremovepileupatthelhcwithjetimages
AT nachmanbenjamin learningtoremovepileupatthelhcwithjetimages
AT schwartzmatthewd learningtoremovepileupatthelhcwithjetimages

Ejemplares similares