Transferability of deep learning models in searches for new physics at colliders
Autor(a) principal: | |
---|---|
Data de Publicação: | 2020 |
Outros Autores: | , , |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/1822/69394 |
Resumo: | In this work we assess the transferability of deep learning models to detect beyond the standard model signals. For this we trained deep neural networks on three different signal models: tZ production via a flavor changing neutral current, pair production of vectorlike T-quarks via standard model gluon fusion and via a heavy gluon decay in a grid of three mass points: 1, 1.2 and 1.4 TeV. These networks were trained with t¯t, Z+jets and dibosons as the main backgrounds. Limits were derived for each signal benchmark using the inference of networks trained on each signal independently, so that we can quantify the degradation of their discriminative power across different signal processes. We determine that the limits are compatible within uncertainties for all networks trained on signals with vectorlike T-quarks, whether they are produced via heavy gluon decay or standard model gluon fusion. The network trained on flavor changing neutral current signal, while struggling the most on the other signals, still produces reasonable limits. These results indicate that deep learning models are capable of providing sensitivity in the search for new physics even if it manifests itself in models not assumed during training. |
id |
RCAP_0803201fdbf230825e7689e0f2d3bed8 |
---|---|
oai_identifier_str |
oai:repositorium.sdum.uminho.pt:1822/69394 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
Transferability of deep learning models in searches for new physics at collidersCiências Naturais::Ciências FísicasScience & TechnologyIn this work we assess the transferability of deep learning models to detect beyond the standard model signals. For this we trained deep neural networks on three different signal models: tZ production via a flavor changing neutral current, pair production of vectorlike T-quarks via standard model gluon fusion and via a heavy gluon decay in a grid of three mass points: 1, 1.2 and 1.4 TeV. These networks were trained with t¯t, Z+jets and dibosons as the main backgrounds. Limits were derived for each signal benchmark using the inference of networks trained on each signal independently, so that we can quantify the degradation of their discriminative power across different signal processes. We determine that the limits are compatible within uncertainties for all networks trained on signals with vectorlike T-quarks, whether they are produced via heavy gluon decay or standard model gluon fusion. The network trained on flavor changing neutral current signal, while struggling the most on the other signals, still produces reasonable limits. These results indicate that deep learning models are capable of providing sensitivity in the search for new physics even if it manifests itself in models not assumed during training.We would like to thank A. Peixoto and J. Santiago foruseful discussions and help with signal generation. We alsoacknowledge the support from FCT Portugal, Lisboa2020,Compete2020, Portugal2020 and FEDER under ProjectNo. PTDC/FIS-PAR/29147/2017 and through GrantNo. PD/BD/135435/2017. The computational part of thiswork was supported by Infraestrutura Nacional deComputação Distribuída (INCD) (funded by FCT andFEDER under Project No. 01/SAICT/2016 n° 022153)and by the Minho Advanced Computing Center (MACC).The Titan Xp GPU card used for the training of the deepneural networks developed for this project was kindlydonated by the NVIDIA Corporation.info:eu-repo/semantics/publishedVersionAmerican Physical Society (APS)Universidade do MinhoRomão, M. CrispimCastro, Nuno FilipePedro, R.Vale, T.20202020-01-01T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/articleapplication/pdfhttp://hdl.handle.net/1822/69394eng2470-00102470-002910.1103/PhysRevD.101.035042https://journals.aps.org/prd/abstract/10.1103/PhysRevD.101.035042info:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2023-07-21T12:52:58Zoai:repositorium.sdum.uminho.pt:1822/69394Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-19T19:52:12.900672Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
Transferability of deep learning models in searches for new physics at colliders |
title |
Transferability of deep learning models in searches for new physics at colliders |
spellingShingle |
Transferability of deep learning models in searches for new physics at colliders Romão, M. Crispim Ciências Naturais::Ciências Físicas Science & Technology |
title_short |
Transferability of deep learning models in searches for new physics at colliders |
title_full |
Transferability of deep learning models in searches for new physics at colliders |
title_fullStr |
Transferability of deep learning models in searches for new physics at colliders |
title_full_unstemmed |
Transferability of deep learning models in searches for new physics at colliders |
title_sort |
Transferability of deep learning models in searches for new physics at colliders |
author |
Romão, M. Crispim |
author_facet |
Romão, M. Crispim Castro, Nuno Filipe Pedro, R. Vale, T. |
author_role |
author |
author2 |
Castro, Nuno Filipe Pedro, R. Vale, T. |
author2_role |
author author author |
dc.contributor.none.fl_str_mv |
Universidade do Minho |
dc.contributor.author.fl_str_mv |
Romão, M. Crispim Castro, Nuno Filipe Pedro, R. Vale, T. |
dc.subject.por.fl_str_mv |
Ciências Naturais::Ciências Físicas Science & Technology |
topic |
Ciências Naturais::Ciências Físicas Science & Technology |
description |
In this work we assess the transferability of deep learning models to detect beyond the standard model signals. For this we trained deep neural networks on three different signal models: tZ production via a flavor changing neutral current, pair production of vectorlike T-quarks via standard model gluon fusion and via a heavy gluon decay in a grid of three mass points: 1, 1.2 and 1.4 TeV. These networks were trained with t¯t, Z+jets and dibosons as the main backgrounds. Limits were derived for each signal benchmark using the inference of networks trained on each signal independently, so that we can quantify the degradation of their discriminative power across different signal processes. We determine that the limits are compatible within uncertainties for all networks trained on signals with vectorlike T-quarks, whether they are produced via heavy gluon decay or standard model gluon fusion. The network trained on flavor changing neutral current signal, while struggling the most on the other signals, still produces reasonable limits. These results indicate that deep learning models are capable of providing sensitivity in the search for new physics even if it manifests itself in models not assumed during training. |
publishDate |
2020 |
dc.date.none.fl_str_mv |
2020 2020-01-01T00:00:00Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/article |
format |
article |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/1822/69394 |
url |
http://hdl.handle.net/1822/69394 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
2470-0010 2470-0029 10.1103/PhysRevD.101.035042 https://journals.aps.org/prd/abstract/10.1103/PhysRevD.101.035042 |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.publisher.none.fl_str_mv |
American Physical Society (APS) |
publisher.none.fl_str_mv |
American Physical Society (APS) |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799133113472778240 |