Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques

CMS Collaboration; Freya Blekman; Emil Sørensen Bols; Simranjit Singh Chhibra; Jorgen D'Hondt; Jarne De Clercq; Denys Lontkovskyi; Steven Lowette; Ivan Marchesini; Seth Moortgat; Quentin Python; Kirill Skovpen; Stefaan Tavernier; Walter Van Doninck; Petra Van Mulders; Douglas Burns; Douglas John Paul Burns

doi:10.1088/1748-0221/15/06/P06005

Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques

CMS Collaboration, Freya Blekman, Emil Sørensen Bols, Simranjit Singh Chhibra, Jorgen D'Hondt, Jarne De Clercq, Denys Lontkovskyi, Steven Lowette, Ivan Marchesini, Seth Moortgat, Quentin Python, Kirill Skovpen, Stefaan Tavernier, Walter Van Doninck, Petra Van Mulders, Douglas Burns, Douglas John Paul Burns

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

79 Citations (Scopus)

101 Downloads (Pure)

Abstract

Machine-learning (ML) techniques are explored to identify and classify hadronic decays of highly Lorentz-boosted W/Z/Higgs bosons and top quarks. Techniques without ML have also been evaluated and are included for comparison. The identification performances of a variety of algorithms are characterized in simulated events and directly compared with data. The algorithms are validated using proton-proton collision data at s = 13TeV, corresponding to an integrated luminosity of 35.9 fb-1. Systematic uncertainties are assessed by comparing the results obtained using simulation and collision data. The new techniques studied in this paper provide significant performance improvements over non-ML techniques, reducing the background rate by up to an order of magnitude at the same signal efficiency.

Original language	English
Article number	P06005
Number of pages	83
Journal	JINST
Volume	15
Issue number	6
DOIs	https://doi.org/10.1088/1748-0221/15/06/P06005
Publication status	Published - 3 Jun 2020

Bibliographical note

Replaced with the published version. Added the journal reference and the DOI. All the figures and tables can be found at http://cms-results.web.cern.ch/cms-results/public-results/publications/JME-18-002 (CMS Public Pages)

Keywords

hep-ex
physics.ins-det

Access to Document

10.1088/1748-0221/15/06/P06005Licence: CC BY

2004.08262v2Submitted manuscript, 5.01 MBLicence: CC BY
Sirunyan_2020_J._Inst._15_P06005Final published version, 5.12 MBLicence: CC BY

Cite this

Collaboration, CMS., Blekman, F., Bols, E. S., Chhibra, S. S., D'Hondt, J., De Clercq, J., Lontkovskyi, D., Lowette, S., Marchesini, I., Moortgat, S., Python, Q., Skovpen, K., Tavernier, S., Van Doninck, W., Van Mulders, P., Burns, D., & Burns, D. J. P. (2020). Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques. JINST, 15(6), [P06005]. https://doi.org/10.1088/1748-0221/15/06/P06005

Collaboration, CMS ; Blekman, Freya ; Bols, Emil Sørensen ; Chhibra, Simranjit Singh ; D'Hondt, Jorgen ; De Clercq, Jarne ; Lontkovskyi, Denys ; Lowette, Steven ; Marchesini, Ivan ; Moortgat, Seth ; Python, Quentin ; Skovpen, Kirill ; Tavernier, Stefaan ; Van Doninck, Walter ; Van Mulders, Petra ; Burns, Douglas ; Burns, Douglas John Paul. / Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques. In: JINST. 2020 ; Vol. 15, No. 6.

@article{2d4037997fa44eda87c57d7f56269cb8,

title = "Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques",

abstract = "Machine-learning (ML) techniques are explored to identify and classify hadronic decays of highly Lorentz-boosted W/Z/Higgs bosons and top quarks. Techniques without ML have also been evaluated and are included for comparison. The identification performances of a variety of algorithms are characterized in simulated events and directly compared with data. The algorithms are validated using proton-proton collision data at s = 13TeV, corresponding to an integrated luminosity of 35.9 fb-1. Systematic uncertainties are assessed by comparing the results obtained using simulation and collision data. The new techniques studied in this paper provide significant performance improvements over non-ML techniques, reducing the background rate by up to an order of magnitude at the same signal efficiency.",

keywords = "hep-ex, physics.ins-det",

author = "CMS Collaboration and Freya Blekman and Bols, {Emil S{\o}rensen} and Chhibra, {Simranjit Singh} and Jorgen D'Hondt and {De Clercq}, Jarne and Denys Lontkovskyi and Steven Lowette and Ivan Marchesini and Seth Moortgat and Quentin Python and Kirill Skovpen and Stefaan Tavernier and {Van Doninck}, Walter and {Van Mulders}, Petra and Douglas Burns and Burns, {Douglas John Paul}",

note = "Replaced with the published version. Added the journal reference and the DOI. All the figures and tables can be found at http://cms-results.web.cern.ch/cms-results/public-results/publications/JME-18-002 (CMS Public Pages)",

year = "2020",

month = jun,

day = "3",

doi = "10.1088/1748-0221/15/06/P06005",

language = "English",

volume = "15",

journal = "Journal of Instrumentation",

issn = "1748-0221",

publisher = "IOP Publishing Ltd.",

number = "6",

}

Collaboration, CMS, Blekman, F, Bols, ES, Chhibra, SS, D'Hondt, J, De Clercq, J, Lontkovskyi, D, Lowette, S, Marchesini, I, Moortgat, S, Python, Q, Skovpen, K, Tavernier, S, Van Doninck, W, Van Mulders, P, Burns, D & Burns, DJP 2020, 'Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques', JINST, vol. 15, no. 6, P06005. https://doi.org/10.1088/1748-0221/15/06/P06005

Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques. / Collaboration, CMS; Blekman, Freya; Bols, Emil Sørensen; Chhibra, Simranjit Singh; D'Hondt, Jorgen; De Clercq, Jarne; Lontkovskyi, Denys; Lowette, Steven; Marchesini, Ivan; Moortgat, Seth; Python, Quentin; Skovpen, Kirill; Tavernier, Stefaan; Van Doninck, Walter; Van Mulders, Petra; Burns, Douglas; Burns, Douglas John Paul.

In: JINST, Vol. 15, No. 6, P06005, 03.06.2020.

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

TY - JOUR

T1 - Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques

AU - Collaboration, CMS

AU - Blekman, Freya

AU - Bols, Emil Sørensen

AU - Chhibra, Simranjit Singh

AU - D'Hondt, Jorgen

AU - De Clercq, Jarne

AU - Lontkovskyi, Denys

AU - Lowette, Steven

AU - Marchesini, Ivan

AU - Moortgat, Seth

AU - Python, Quentin

AU - Skovpen, Kirill

AU - Tavernier, Stefaan

AU - Van Doninck, Walter

AU - Van Mulders, Petra

AU - Burns, Douglas

AU - Burns, Douglas John Paul

N1 - Replaced with the published version. Added the journal reference and the DOI. All the figures and tables can be found at http://cms-results.web.cern.ch/cms-results/public-results/publications/JME-18-002 (CMS Public Pages)

PY - 2020/6/3

Y1 - 2020/6/3

N2 - Machine-learning (ML) techniques are explored to identify and classify hadronic decays of highly Lorentz-boosted W/Z/Higgs bosons and top quarks. Techniques without ML have also been evaluated and are included for comparison. The identification performances of a variety of algorithms are characterized in simulated events and directly compared with data. The algorithms are validated using proton-proton collision data at s = 13TeV, corresponding to an integrated luminosity of 35.9 fb-1. Systematic uncertainties are assessed by comparing the results obtained using simulation and collision data. The new techniques studied in this paper provide significant performance improvements over non-ML techniques, reducing the background rate by up to an order of magnitude at the same signal efficiency.

AB - Machine-learning (ML) techniques are explored to identify and classify hadronic decays of highly Lorentz-boosted W/Z/Higgs bosons and top quarks. Techniques without ML have also been evaluated and are included for comparison. The identification performances of a variety of algorithms are characterized in simulated events and directly compared with data. The algorithms are validated using proton-proton collision data at s = 13TeV, corresponding to an integrated luminosity of 35.9 fb-1. Systematic uncertainties are assessed by comparing the results obtained using simulation and collision data. The new techniques studied in this paper provide significant performance improvements over non-ML techniques, reducing the background rate by up to an order of magnitude at the same signal efficiency.

KW - hep-ex

KW - physics.ins-det

UR - http://www.scopus.com/inward/record.url?scp=85088524436&partnerID=8YFLogxK

U2 - 10.1088/1748-0221/15/06/P06005

DO - 10.1088/1748-0221/15/06/P06005

M3 - Article

VL - 15

JO - Journal of Instrumentation

JF - Journal of Instrumentation

SN - 1748-0221

IS - 6

M1 - P06005

ER -

Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

SRP8: Strategic Research Programme: High-Energy Physics at the VUB

Cite this

Identification of heavy, energetic, hadronically decaying particles using machine-learning techniques

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Projects

SRP8: Strategic Research Programme: High-Energy Physics at the VUB

Cite this