Publicação

Convolutional neural networks for the classification of glitches in gravitational-wave data streams

Ver documento

Detalhes bibliográficos
Resumo:We investigate the use of convolutional neural networks (including the modern ConvNeXt network family) to classify transient noise signals (i.e. glitches) and gravitational waves (GWs) in data from the Advanced LIGO detectors. First, we use models with a supervised learning approach, both trained from scratch using the Gravity Spy dataset and employing transfer learning by fine-tuning pre-trained models in this dataset. Second, we also explore a self-supervised approach, pre-training models with automatically generated pseudo-labels. Our findings are very close to existing results for the same dataset, reaching values for the F1 score of 97.18% (94.15%) for the best supervised (self-supervised) model. We further test the models using actual GW signals from LIGO-Virgo's O3 run. Although trained using data from previous runs (O1 and O2), the models show good performance, in particular when using transfer learning. We find that transfer learning improves the scores without the need for any training on real signals apart from the less than 50 chirp examples from hardware injections present in the Gravity Spy dataset. This motivates the use of transfer learning not only for glitch classification but also for signal classification.
Autores principais:Fernandes, Tiago
Outros Autores:Vieira, Samuel; Onofre, A.; Calderón Bustillo, Juan; Torres-Forné, Alejandro; Font, José A.
Assunto:deep learning glitches gravitational waves machine learning
Ano:2023
País:Portugal
Tipo de documento:artigo
Tipo de acesso:acesso restrito
Instituição associada:Universidade do Minho
Idioma:inglês
Origem:RepositóriUM - Universidade do Minho
_version_ 1866270943256510464
author Fernandes, Tiago
author2 Vieira, Samuel
Onofre, A.
Calderón Bustillo, Juan
Torres-Forné, Alejandro
Font, José A.
author2_role author
author
author
author
author
author_facet Fernandes, Tiago
Vieira, Samuel
Onofre, A.
Calderón Bustillo, Juan
Torres-Forné, Alejandro
Font, José A.
author_role author
contributor_name_str_mv Universidade do Minho
country_str PT
creators_json_txt [{\"Person.name\":\"Fernandes, Tiago\"},{\"Person.name\":\"Vieira, Samuel\"},{\"Person.name\":\"Onofre, A.\"},{\"Person.name\":\"Calderón Bustillo, Juan\"},{\"Person.name\":\"Torres-Forné, Alejandro\"},{\"Person.name\":\"Font, José A.\"}]
datacite.contributors.contributor.contributorName.fl_str_mv Universidade do Minho
datacite.creators.creator.creatorName.fl_str_mv Fernandes, Tiago
Vieira, Samuel
Onofre, A.
Calderón Bustillo, Juan
Torres-Forné, Alejandro
Font, José A.
datacite.date.Accepted.fl_str_mv 2023-01-01T00:00:00Z
datacite.date.available.fl_str_mv 2024-01-02T21:57:15Z
datacite.date.embargoed.fl_str_mv 2024-01-02T21:57:15Z
datacite.rights.fl_str_mv http://purl.org/coar/access_right/c_16ec
datacite.subjects.subject.fl_str_mv deep learning
glitches
gravitational waves
machine learning
datacite.titles.title.fl_str_mv Convolutional neural networks for the classification of glitches in gravitational-wave data streams
dc.contributor.none.fl_str_mv Universidade do Minho
dc.creator.none.fl_str_mv Fernandes, Tiago
Vieira, Samuel
Onofre, A.
Calderón Bustillo, Juan
Torres-Forné, Alejandro
Font, José A.
dc.date.Accepted.fl_str_mv 2023-01-01T00:00:00Z
dc.date.available.fl_str_mv 2024-01-02T21:57:15Z
dc.date.embargoed.fl_str_mv 2024-01-02T21:57:15Z
dc.format.none.fl_str_mv application/pdf
application/pdf
dc.identifier.none.fl_str_mv https://hdl.handle.net/1822/87778
dc.language.none.fl_str_mv eng
dc.publisher.none.fl_str_mv IOP Publishing
dc.rights.none.fl_str_mv http://purl.org/coar/access_right/c_16ec
dc.subject.none.fl_str_mv deep learning
glitches
gravitational waves
machine learning
dc.title.fl_str_mv Convolutional neural networks for the classification of glitches in gravitational-wave data streams
dc.type.none.fl_str_mv http://purl.org/coar/resource_type/c_6501
description We investigate the use of convolutional neural networks (including the modern ConvNeXt network family) to classify transient noise signals (i.e. glitches) and gravitational waves (GWs) in data from the Advanced LIGO detectors. First, we use models with a supervised learning approach, both trained from scratch using the Gravity Spy dataset and employing transfer learning by fine-tuning pre-trained models in this dataset. Second, we also explore a self-supervised approach, pre-training models with automatically generated pseudo-labels. Our findings are very close to existing results for the same dataset, reaching values for the F1 score of 97.18% (94.15%) for the best supervised (self-supervised) model. We further test the models using actual GW signals from LIGO-Virgo's O3 run. Although trained using data from previous runs (O1 and O2), the models show good performance, in particular when using transfer learning. We find that transfer learning improves the scores without the need for any training on real signals apart from the less than 50 chirp examples from hardware injections present in the Gravity Spy dataset. This motivates the use of transfer learning not only for glitch classification but also for signal classification.
dirty 0
eu_rights_str_mv restrictedAccess
format article
fulltext.url.fl_str_mv https://prod-dspace.uminho.pt/bitstreams/fc7b6bc1-c006-41de-af2f-ae4c72773f35/download
id rum_d2fcc8958a23cd9eb2b0d820069f1e21
identifier.url.fl_str_mv https://hdl.handle.net/1822/87778
instacron_str repositorium
institution Universidade do Minho
instname_str Universidade do Minho
language eng
network_acronym_str rum
network_name_str RepositóriUM - Universidade do Minho
oai_identifier_str oai:repositorium.uminho.pt:1822/87778
organization_str_mv urn:organizationAcronym:repositorium
person_str_mv Fernandes, Tiago
Vieira, Samuel
Onofre, A.
Calderón Bustillo, Juan
Torres-Forné, Alejandro
Font, José A.
publishDate 2023
publisher.none.fl_str_mv IOP Publishing
reponame_str RepositóriUM - Universidade do Minho
repository_id_str urn:repositoryAcronym:rum
service_str_mv urn:repositoryAcronym:rum
spelling engIOP PublishingporWe investigate the use of convolutional neural networks (including the modern ConvNeXt network family) to classify transient noise signals (i.e. glitches) and gravitational waves (GWs) in data from the Advanced LIGO detectors. First, we use models with a supervised learning approach, both trained from scratch using the Gravity Spy dataset and employing transfer learning by fine-tuning pre-trained models in this dataset. Second, we also explore a self-supervised approach, pre-training models with automatically generated pseudo-labels. Our findings are very close to existing results for the same dataset, reaching values for the F1 score of 97.18% (94.15%) for the best supervised (self-supervised) model. We further test the models using actual GW signals from LIGO-Virgo's O3 run. Although trained using data from previous runs (O1 and O2), the models show good performance, in particular when using transfer learning. We find that transfer learning improves the scores without the need for any training on real signals apart from the less than 50 chirp examples from hardware injections present in the Gravity Spy dataset. This motivates the use of transfer learning not only for glitch classification but also for signal classification.application/pdfapplication/pdfporConvolutional neural networks for the classification of glitches in gravitational-wave data streamsFernandes, TiagoVieira, SamuelOnofre, A.Calderón Bustillo, JuanTorres-Forné, AlejandroFont, José A.HostingInstitutionOrganizationalUniversidade do Minhoe-mailmailto:repositorium@usdb.uminho.ptrepositorium@usdb.uminho.ptISSNIsPartOf0264-9381DOIIsPartOf10.1088/1361-6382/acf26c2024-01-02T21:57:15Z20232023-01-01T00:00:00ZHandlehttps://hdl.handle.net/1822/87778http://purl.org/coar/access_right/c_16ecrestricted accessdeep learningglitchesgravitational wavesmachine learning1724761 bytes2078004 bytesliteraturehttp://purl.org/coar/resource_type/c_6501journal articlehttp://purl.org/coar/access_right/c_16ecapplication/pdffulltexthttps://prod-dspace.uminho.pt/bitstreams/fc7b6bc1-c006-41de-af2f-ae4c72773f35/downloadhttp://purl.org/coar/access_right/c_abf2application/pdffulltexthttps://prod-dspace.uminho.pt/bitstreams/bc90ab64-ccbc-4151-9778-4b1bf9b618ed/download
spellingShingle Convolutional neural networks for the classification of glitches in gravitational-wave data streams
Fernandes, Tiago
deep learning
glitches
gravitational waves
machine learning
status SINGLETON
subject.fl_str_mv deep learning
glitches
gravitational waves
machine learning
title Convolutional neural networks for the classification of glitches in gravitational-wave data streams
title_full Convolutional neural networks for the classification of glitches in gravitational-wave data streams
title_fullStr Convolutional neural networks for the classification of glitches in gravitational-wave data streams
title_full_unstemmed Convolutional neural networks for the classification of glitches in gravitational-wave data streams
title_short Convolutional neural networks for the classification of glitches in gravitational-wave data streams
title_sort Convolutional neural networks for the classification of glitches in gravitational-wave data streams
topic deep learning
glitches
gravitational waves
machine learning
topic_facet deep learning
glitches
gravitational waves
machine learning
url https://hdl.handle.net/1822/87778
visible 1