ℓ<sub>1</sub>-norm based safe semi-supervised learning

In the past few years, Safe Semi-Supervised Learning (S3L) has received considerable attentions in machine learning field. Different researchers have proposed many S3L methods for safe exploitation of risky unlabeled samples which result in performance degradation of Semi-Supervised Learning (SSL)....

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Haitao Gan, Zhi Yang, Ji Wang, Bing Li
Formato:	article
Lenguaje:	EN
Publicado:	AIMS Press 2021
Materias:	semi-supervised learning safe semi-supervised learning performance degradation ℓ<sub>1</sub> norm importance estimation Biotechnology TP248.13-248.65 Mathematics QA1-939
Acceso en línea:	https://doaj.org/article/0c8c3b2da2574d8788692f62e02d3bcb
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:0c8c3b2da2574d8788692f62e02d3bcb
record_format	dspace
spelling	oai:doaj.org-article:0c8c3b2da2574d8788692f62e02d3bcb2021-11-23T02:36:33Zℓ<sub>1</sub>-norm based safe semi-supervised learning10.3934/mbe.20213831551-0018https://doaj.org/article/0c8c3b2da2574d8788692f62e02d3bcb2021-09-01T00:00:00Zhttps://www.aimspress.com/article/doi/10.3934/mbe.2021383?viewType=HTMLhttps://doaj.org/toc/1551-0018In the past few years, Safe Semi-Supervised Learning (S3L) has received considerable attentions in machine learning field. Different researchers have proposed many S3L methods for safe exploitation of risky unlabeled samples which result in performance degradation of Semi-Supervised Learning (SSL). Nevertheless, there exist some shortcomings: (1) Risk degrees of the unlabeled samples are in advance defined by analyzing prediction differences between Supervised Learning (SL) and SSL; (2) Negative impacts of labeled samples on learning performance are not investigated. Therefore, it is essential to design a novel method to adaptively estimate importance and risk of both unlabeled and labeled samples. For this purpose, we present $ \ell_{1} $-norm based S3L which can simultaneously reach the safe exploitation of the labeled and unlabeled samples in this paper. In order to solve the proposed ptimization problem, we utilize an effective iterative approach. In each iteration, one can adaptively estimate the weights of both labeled and unlabeled samples. The weights can reflect the importance or risk of the labeled and unlabeled samples. Hence, the negative effects of the labeled and unlabeled samples are expected to be reduced. Experimental performance on different datasets verifies that the proposed S3L method can obtain comparable performance with the existing SL, SSL and S3L methods and achieve the expected goal.Haitao GanZhi Yang Ji Wang Bing LiAIMS Pressarticlesemi-supervised learningsafe semi-supervised learningperformance degradationℓ<sub>1</sub> normimportance estimationBiotechnologyTP248.13-248.65MathematicsQA1-939ENMathematical Biosciences and Engineering, Vol 18, Iss 6, Pp 7727-7742 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	semi-supervised learning safe semi-supervised learning performance degradation ℓ<sub>1</sub> norm importance estimation Biotechnology TP248.13-248.65 Mathematics QA1-939
spellingShingle	semi-supervised learning safe semi-supervised learning performance degradation ℓ<sub>1</sub> norm importance estimation Biotechnology TP248.13-248.65 Mathematics QA1-939 Haitao Gan Zhi Yang Ji Wang Bing Li ℓ<sub>1</sub>-norm based safe semi-supervised learning
description	In the past few years, Safe Semi-Supervised Learning (S3L) has received considerable attentions in machine learning field. Different researchers have proposed many S3L methods for safe exploitation of risky unlabeled samples which result in performance degradation of Semi-Supervised Learning (SSL). Nevertheless, there exist some shortcomings: (1) Risk degrees of the unlabeled samples are in advance defined by analyzing prediction differences between Supervised Learning (SL) and SSL; (2) Negative impacts of labeled samples on learning performance are not investigated. Therefore, it is essential to design a novel method to adaptively estimate importance and risk of both unlabeled and labeled samples. For this purpose, we present $ \ell_{1} $-norm based S3L which can simultaneously reach the safe exploitation of the labeled and unlabeled samples in this paper. In order to solve the proposed ptimization problem, we utilize an effective iterative approach. In each iteration, one can adaptively estimate the weights of both labeled and unlabeled samples. The weights can reflect the importance or risk of the labeled and unlabeled samples. Hence, the negative effects of the labeled and unlabeled samples are expected to be reduced. Experimental performance on different datasets verifies that the proposed S3L method can obtain comparable performance with the existing SL, SSL and S3L methods and achieve the expected goal.
format	article
author	Haitao Gan Zhi Yang Ji Wang Bing Li
author_facet	Haitao Gan Zhi Yang Ji Wang Bing Li
author_sort	Haitao Gan
title	ℓ<sub>1</sub>-norm based safe semi-supervised learning
title_short	ℓ<sub>1</sub>-norm based safe semi-supervised learning
title_full	ℓ<sub>1</sub>-norm based safe semi-supervised learning
title_fullStr	ℓ<sub>1</sub>-norm based safe semi-supervised learning
title_full_unstemmed	ℓ<sub>1</sub>-norm based safe semi-supervised learning
title_sort	ℓ<sub>1</sub>-norm based safe semi-supervised learning
publisher	AIMS Press
publishDate	2021
url	https://doaj.org/article/0c8c3b2da2574d8788692f62e02d3bcb
work_keys_str_mv	AT haitaogan lsub1subnormbasedsafesemisupervisedlearning AT zhiyang lsub1subnormbasedsafesemisupervisedlearning AT jiwang lsub1subnormbasedsafesemisupervisedlearning AT bingli lsub1subnormbasedsafesemisupervisedlearning
_version_	1718417360724951040

ℓ<sub>1</sub>-norm based safe semi-supervised learning

Ejemplares similares