Weakly supervised learning based on hypergraph manifold ranking
Autor(a) principal: | |
---|---|
Data de Publicação: | 2022 |
Outros Autores: | , , , , , |
Tipo de documento: | Artigo |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UNESP |
Texto Completo: | http://dx.doi.org/10.1016/j.jvcir.2022.103666 http://hdl.handle.net/11449/249306 |
Resumo: | Significant challenges still remain despite the impressive recent advances in machine learning techniques, particularly in multimedia data understanding. One of the main challenges in real-world scenarios is the nature and relation between training and test datasets. Very often, only small sets of coarse-grained labeled data are available to train models, which are expected to be applied on large datasets and fine-grained tasks. Weakly supervised learning approaches handle such constraints by maximizing useful training information in labeled and unlabeled data. In this research direction, we propose a weakly supervised approach that analyzes the dataset manifold to expand the available labeled set. A hypergraph manifold ranking algorithm is exploited to represent the contextual similarity information encoded in the unlabeled data and identify strong similarity relations, which are taken as a path to label expansion. The expanded labeled set is subsequently exploited for a more comprehensive and accurate training process. The proposed model was evaluated jointly with supervised and semi-supervised classifiers, including Graph Convolutional Networks. The experimental results on image and video datasets demonstrate significant gains and accurate results for different classifiers in diverse scenarios. |
id |
UNSP_4dfb0c57de9701bde0f75e1a250802ad |
---|---|
oai_identifier_str |
oai:repositorio.unesp.br:11449/249306 |
network_acronym_str |
UNSP |
network_name_str |
Repositório Institucional da UNESP |
repository_id_str |
2946 |
spelling |
Weakly supervised learning based on hypergraph manifold rankingHypergraphManifold learningRankingWeakly supervised learningSignificant challenges still remain despite the impressive recent advances in machine learning techniques, particularly in multimedia data understanding. One of the main challenges in real-world scenarios is the nature and relation between training and test datasets. Very often, only small sets of coarse-grained labeled data are available to train models, which are expected to be applied on large datasets and fine-grained tasks. Weakly supervised learning approaches handle such constraints by maximizing useful training information in labeled and unlabeled data. In this research direction, we propose a weakly supervised approach that analyzes the dataset manifold to expand the available labeled set. A hypergraph manifold ranking algorithm is exploited to represent the contextual similarity information encoded in the unlabeled data and identify strong similarity relations, which are taken as a path to label expansion. The expanded labeled set is subsequently exploited for a more comprehensive and accurate training process. The proposed model was evaluated jointly with supervised and semi-supervised classifiers, including Graph Convolutional Networks. The experimental results on image and video datasets demonstrate significant gains and accurate results for different classifiers in diverse scenarios.Microsoft ResearchPetrobrasFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)Department of Statistics Applied Mathematics and Computing State University of São Paulo (UNESP), Av. 24-A, 1515Institute of Science and Technology Federal University of São Paulo (UNIFESP)School of Sciences State University of São Paulo (UNESP)Department of Computing Federal University of São Carlos (UFSCAR)Department of Statistics Applied Mathematics and Computing State University of São Paulo (UNESP), Av. 24-A, 1515School of Sciences State University of São Paulo (UNESP)Petrobras: #2017/ 00285-6FAPESP: #2017/25908-6FAPESP: #2018/15597-6FAPESP: #2018/23908-1FAPESP: #2019/ 04754-6FAPESP: #2020/11366-0CNPq: #309439/2020-5CNPq: #314868/2020-8CNPq: #422667/2021-8Universidade Estadual Paulista (UNESP)Universidade de São Paulo (USP)Universidade Federal de São Carlos (UFSCar)Presotto, João Gabriel Camacho [UNESP]dos Santos, Samuel FelipeValem, Lucas Pascotti [UNESP]Faria, Fabio AugustoPapa, João Paulo [UNESP]Almeida, JurandyPedronette, Daniel Carlos Guimarães [UNESP]2023-07-29T15:12:29Z2023-07-29T15:12:29Z2022-11-01info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/articlehttp://dx.doi.org/10.1016/j.jvcir.2022.103666Journal of Visual Communication and Image Representation, v. 89.1095-90761047-3203http://hdl.handle.net/11449/24930610.1016/j.jvcir.2022.1036662-s2.0-85140708774Scopusreponame:Repositório Institucional da UNESPinstname:Universidade Estadual Paulista (UNESP)instacron:UNESPengJournal of Visual Communication and Image Representationinfo:eu-repo/semantics/openAccess2024-04-23T16:11:00Zoai:repositorio.unesp.br:11449/249306Repositório InstitucionalPUBhttp://repositorio.unesp.br/oai/requestopendoar:29462024-08-05T23:01:31.942691Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP)false |
dc.title.none.fl_str_mv |
Weakly supervised learning based on hypergraph manifold ranking |
title |
Weakly supervised learning based on hypergraph manifold ranking |
spellingShingle |
Weakly supervised learning based on hypergraph manifold ranking Presotto, João Gabriel Camacho [UNESP] Hypergraph Manifold learning Ranking Weakly supervised learning |
title_short |
Weakly supervised learning based on hypergraph manifold ranking |
title_full |
Weakly supervised learning based on hypergraph manifold ranking |
title_fullStr |
Weakly supervised learning based on hypergraph manifold ranking |
title_full_unstemmed |
Weakly supervised learning based on hypergraph manifold ranking |
title_sort |
Weakly supervised learning based on hypergraph manifold ranking |
author |
Presotto, João Gabriel Camacho [UNESP] |
author_facet |
Presotto, João Gabriel Camacho [UNESP] dos Santos, Samuel Felipe Valem, Lucas Pascotti [UNESP] Faria, Fabio Augusto Papa, João Paulo [UNESP] Almeida, Jurandy Pedronette, Daniel Carlos Guimarães [UNESP] |
author_role |
author |
author2 |
dos Santos, Samuel Felipe Valem, Lucas Pascotti [UNESP] Faria, Fabio Augusto Papa, João Paulo [UNESP] Almeida, Jurandy Pedronette, Daniel Carlos Guimarães [UNESP] |
author2_role |
author author author author author author |
dc.contributor.none.fl_str_mv |
Universidade Estadual Paulista (UNESP) Universidade de São Paulo (USP) Universidade Federal de São Carlos (UFSCar) |
dc.contributor.author.fl_str_mv |
Presotto, João Gabriel Camacho [UNESP] dos Santos, Samuel Felipe Valem, Lucas Pascotti [UNESP] Faria, Fabio Augusto Papa, João Paulo [UNESP] Almeida, Jurandy Pedronette, Daniel Carlos Guimarães [UNESP] |
dc.subject.por.fl_str_mv |
Hypergraph Manifold learning Ranking Weakly supervised learning |
topic |
Hypergraph Manifold learning Ranking Weakly supervised learning |
description |
Significant challenges still remain despite the impressive recent advances in machine learning techniques, particularly in multimedia data understanding. One of the main challenges in real-world scenarios is the nature and relation between training and test datasets. Very often, only small sets of coarse-grained labeled data are available to train models, which are expected to be applied on large datasets and fine-grained tasks. Weakly supervised learning approaches handle such constraints by maximizing useful training information in labeled and unlabeled data. In this research direction, we propose a weakly supervised approach that analyzes the dataset manifold to expand the available labeled set. A hypergraph manifold ranking algorithm is exploited to represent the contextual similarity information encoded in the unlabeled data and identify strong similarity relations, which are taken as a path to label expansion. The expanded labeled set is subsequently exploited for a more comprehensive and accurate training process. The proposed model was evaluated jointly with supervised and semi-supervised classifiers, including Graph Convolutional Networks. The experimental results on image and video datasets demonstrate significant gains and accurate results for different classifiers in diverse scenarios. |
publishDate |
2022 |
dc.date.none.fl_str_mv |
2022-11-01 2023-07-29T15:12:29Z 2023-07-29T15:12:29Z |
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://dx.doi.org/10.1016/j.jvcir.2022.103666 Journal of Visual Communication and Image Representation, v. 89. 1095-9076 1047-3203 http://hdl.handle.net/11449/249306 10.1016/j.jvcir.2022.103666 2-s2.0-85140708774 |
url |
http://dx.doi.org/10.1016/j.jvcir.2022.103666 http://hdl.handle.net/11449/249306 |
identifier_str_mv |
Journal of Visual Communication and Image Representation, v. 89. 1095-9076 1047-3203 10.1016/j.jvcir.2022.103666 2-s2.0-85140708774 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.relation.none.fl_str_mv |
Journal of Visual Communication and Image Representation |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.source.none.fl_str_mv |
Scopus reponame:Repositório Institucional da UNESP instname:Universidade Estadual Paulista (UNESP) instacron:UNESP |
instname_str |
Universidade Estadual Paulista (UNESP) |
instacron_str |
UNESP |
institution |
UNESP |
reponame_str |
Repositório Institucional da UNESP |
collection |
Repositório Institucional da UNESP |
repository.name.fl_str_mv |
Repositório Institucional da UNESP - Universidade Estadual Paulista (UNESP) |
repository.mail.fl_str_mv |
|
_version_ |
1808129483644862464 |