Probing Higgs boson exotic decays at the LHC with machine learning

Sunghoon Jung; Zhen Liu; Lian Tao Wang; Ke Pan Xie

doi:10.1103/PhysRevD.105.035008

Probing Higgs boson exotic decays at the LHC with machine learning

Sunghoon Jung
, Zhen Liu
, Lian Tao Wang
, Ke Pan Xie

Physics and Astronomy (Twin Cities)

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

8 Scopus citations

Abstract

We study the tagging of Higgs exotic decay signals using different types of deep neural networks (DNNs) while focusing on the W±h associated production channel followed by Higgs decaying into n b quarks with n=4, 6, and 8. All the Higgs decay products are collected into a fat jet, to which we apply further selection using the DNNs. Three kinds of DNNs are considered - namely, the convolutional neural network, the recursive neural network, and the particle flow network (PFN). The PFN can achieve the best performance because its structure allows one to enfold more information in addition to the four-momenta of the jet constituents, such as the particle identifier, and tracks the parameters. Using the PFN as an example, we verify that it can serve as an efficient tagger even though it is trained on a different event topology with different b multiplicity from the actual signal. The projected sensitivity to the branching ratio of Higgs decaying into n b quarks at the HL-LHC are 10%, 3%, and 1%, for n=4, 6, and 8, respectively.

Original language	English (US)
Article number	035008
Journal	Physical Review D
Volume	105
Issue number	3
DOIs	https://doi.org/10.1103/PhysRevD.105.035008
State	Published - Feb 1 2022

Bibliographical note

Funding Information:
We thank Taoli Cheng, Yandong Liu, Gilles Louppe, Yongcheng Wu, Daneng Yang, and Rui Zhang for the useful discussions. We are especially grateful to Gilles Louppe for communication on RecNN and for sharing the datasets. K.-P. X. would like to thank Guang-Ze Fu for the great help on coding. The work of Z. L. is supported in part by the U.S. Department of Energy under Grant No. DE-SC0022345. S. J. and K.-P. X. are supported by Grant Korea No. NRF-2019R1C1C1010050. S. J. is also supported by a POSCO Science Fellowship, and K.-P. X. by the University of Nebraska–Lincoln. K.-P. X. would like to thank the University of Chicago, where part of this work was performed, for the hospitality. Z. L. and L.-T. W. acknowledge the Aspen Center of Physics for its hospitality during the final phase of this study, which is supported by National Science Foundation Grant No. PHY-1607611.

Publisher Copyright:
© 2022 authors. Published by the American Physical Society.

Publisher link

10.1103/PhysRevD.105.035008

Find It

Find It at U of M Libraries

Cite this

@article{67ce712b39364dcea29e3339282d0e90,

title = "Probing Higgs boson exotic decays at the LHC with machine learning",

abstract = "We study the tagging of Higgs exotic decay signals using different types of deep neural networks (DNNs) while focusing on the W±h associated production channel followed by Higgs decaying into n b quarks with n=4, 6, and 8. All the Higgs decay products are collected into a fat jet, to which we apply further selection using the DNNs. Three kinds of DNNs are considered - namely, the convolutional neural network, the recursive neural network, and the particle flow network (PFN). The PFN can achieve the best performance because its structure allows one to enfold more information in addition to the four-momenta of the jet constituents, such as the particle identifier, and tracks the parameters. Using the PFN as an example, we verify that it can serve as an efficient tagger even though it is trained on a different event topology with different b multiplicity from the actual signal. The projected sensitivity to the branching ratio of Higgs decaying into n b quarks at the HL-LHC are 10\%, 3\%, and 1\%, for n=4, 6, and 8, respectively.",

author = "Sunghoon Jung and Zhen Liu and Wang, \{Lian Tao\} and Xie, \{Ke Pan\}",

note = "Funding Information: We thank Taoli Cheng, Yandong Liu, Gilles Louppe, Yongcheng Wu, Daneng Yang, and Rui Zhang for the useful discussions. We are especially grateful to Gilles Louppe for communication on RecNN and for sharing the datasets. K.-P. X. would like to thank Guang-Ze Fu for the great help on coding. The work of Z. L. is supported in part by the U.S. Department of Energy under Grant No. DE-SC0022345. S. J. and K.-P. X. are supported by Grant Korea No. NRF-2019R1C1C1010050. S. J. is also supported by a POSCO Science Fellowship, and K.-P. X. by the University of Nebraska–Lincoln. K.-P. X. would like to thank the University of Chicago, where part of this work was performed, for the hospitality. Z. L. and L.-T. W. acknowledge the Aspen Center of Physics for its hospitality during the final phase of this study, which is supported by National Science Foundation Grant No. PHY-1607611. Publisher Copyright: {\textcopyright} 2022 authors. Published by the American Physical Society.",

year = "2022",

month = feb,

day = "1",

doi = "10.1103/PhysRevD.105.035008",

language = "English (US)",

volume = "105",

journal = "Physical Review D",

issn = "2470-0010",

publisher = "American Physical Society",

number = "3",

}

TY - JOUR

T1 - Probing Higgs boson exotic decays at the LHC with machine learning

AU - Jung, Sunghoon

AU - Liu, Zhen

AU - Wang, Lian Tao

AU - Xie, Ke Pan

N1 - Funding Information: We thank Taoli Cheng, Yandong Liu, Gilles Louppe, Yongcheng Wu, Daneng Yang, and Rui Zhang for the useful discussions. We are especially grateful to Gilles Louppe for communication on RecNN and for sharing the datasets. K.-P. X. would like to thank Guang-Ze Fu for the great help on coding. The work of Z. L. is supported in part by the U.S. Department of Energy under Grant No. DE-SC0022345. S. J. and K.-P. X. are supported by Grant Korea No. NRF-2019R1C1C1010050. S. J. is also supported by a POSCO Science Fellowship, and K.-P. X. by the University of Nebraska–Lincoln. K.-P. X. would like to thank the University of Chicago, where part of this work was performed, for the hospitality. Z. L. and L.-T. W. acknowledge the Aspen Center of Physics for its hospitality during the final phase of this study, which is supported by National Science Foundation Grant No. PHY-1607611. Publisher Copyright: © 2022 authors. Published by the American Physical Society.

PY - 2022/2/1

Y1 - 2022/2/1

N2 - We study the tagging of Higgs exotic decay signals using different types of deep neural networks (DNNs) while focusing on the W±h associated production channel followed by Higgs decaying into n b quarks with n=4, 6, and 8. All the Higgs decay products are collected into a fat jet, to which we apply further selection using the DNNs. Three kinds of DNNs are considered - namely, the convolutional neural network, the recursive neural network, and the particle flow network (PFN). The PFN can achieve the best performance because its structure allows one to enfold more information in addition to the four-momenta of the jet constituents, such as the particle identifier, and tracks the parameters. Using the PFN as an example, we verify that it can serve as an efficient tagger even though it is trained on a different event topology with different b multiplicity from the actual signal. The projected sensitivity to the branching ratio of Higgs decaying into n b quarks at the HL-LHC are 10%, 3%, and 1%, for n=4, 6, and 8, respectively.

AB - We study the tagging of Higgs exotic decay signals using different types of deep neural networks (DNNs) while focusing on the W±h associated production channel followed by Higgs decaying into n b quarks with n=4, 6, and 8. All the Higgs decay products are collected into a fat jet, to which we apply further selection using the DNNs. Three kinds of DNNs are considered - namely, the convolutional neural network, the recursive neural network, and the particle flow network (PFN). The PFN can achieve the best performance because its structure allows one to enfold more information in addition to the four-momenta of the jet constituents, such as the particle identifier, and tracks the parameters. Using the PFN as an example, we verify that it can serve as an efficient tagger even though it is trained on a different event topology with different b multiplicity from the actual signal. The projected sensitivity to the branching ratio of Higgs decaying into n b quarks at the HL-LHC are 10%, 3%, and 1%, for n=4, 6, and 8, respectively.

UR - https://www.scopus.com/pages/publications/85125418255

UR - https://www.scopus.com/pages/publications/85125418255#tab=citedBy

U2 - 10.1103/PhysRevD.105.035008

DO - 10.1103/PhysRevD.105.035008

M3 - Article

AN - SCOPUS:85125418255

SN - 2470-0010

VL - 105

JO - Physical Review D

JF - Physical Review D

IS - 3

M1 - 035008

ER -

Probing Higgs boson exotic decays at the LHC with machine learning

Abstract

Bibliographical note

Publisher link

Find It

Other files and links

Fingerprint

Cite this