Weakly Supervised Learning through Rank-based Contextual Measures

被引:3
作者
Camacho Presotto, Joao Gabriel [1 ]
Valem, Lucas Pascotti [1 ]
de Sa, Nikolas Gomes [1 ]
Guimaraes Pedronette, Daniel Carlos [1 ]
Papa, Joao Paulo [2 ]
机构
[1] UNESP Sao Paulo State Univ, Dept Stat Appl Math & Comp, Rio Claro, SP, Brazil
[2] UNESP Sao Paulo State Univ, Sch Sci, Bauru, SP, Brazil
来源
2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR) | 2021年
基金
巴西圣保罗研究基金会;
关键词
classification; machine learning; weak supervision; rank correlation measure;
D O I
10.1109/ICPR48806.2021.9412596
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Machine learning approaches have achieved remarkable advances over the last decades, especially in supervised learning tasks such as classification. Meanwhile, multimedia data and applications experienced an explosive growth, becoming ubiquitous in diverse domains. Due to the huge increase in multimedia data collections and the lack of labeled data in several scenarios, creating methods capable of exploiting the unlabeled data and operating under weakly supervision is imperative. In this work, we propose a rank-based model to exploit contextual information encoded in the unlabeled data in order to perform weakly supervised classification. We employ different rank-based correlation measures for identifying strong similarities relationships and expanding the labeled set in an unsupervised way. Subsequently, the extended labeled set is used by a classifier to achieve better accuracy results. The proposed weakly supervised approach was evaluated on multimedia classification tasks, considering several combinations of rank correlation measures and classifiers. An experimental evaluation was conducted on 4 public image datasets and different features. Very positive gains were achieved in comparison with various semi-supervised and supervised classifiers taken as baselines when considering the same amount of labeled data.
引用
收藏
页码:5752 / 5759
页数:8
相关论文
共 32 条
[1]   Regularized Diffusion Process on Bidirectional Context for Object Retrieval [J].
Bai, Song ;
Bai, Xiang ;
Tian, Qi ;
Latecki, Longin Jan .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2019, 41 (05) :1213-1226
[2]  
Chapelle O., 2006, Mit Pr, V2006
[3]   Not-so-supervised: A survey of semi-supervised, multi-instance, and transfer learning in medical image analysis [J].
Cheplygina, Veronika ;
de Bruijne, Marleen ;
Pluim, Josien P. W. .
MEDICAL IMAGE ANALYSIS, 2019, 54 :280-296
[4]   SUPPORT-VECTOR NETWORKS [J].
CORTES, C ;
VAPNIK, V .
MACHINE LEARNING, 1995, 20 (03) :273-297
[5]   Diffusion Processes for Retrieval Revisited [J].
Donoser, Michael ;
Bischof, Horst .
2013 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2013, :1320-1327
[6]  
Fagin R, 2003, SIAM PROC S, P28
[7]  
Franceschi L, 2019, PR MACH LEARN RES, V97
[8]   Multimedia Retrieval Through Unsupervised Hypergraph-Based Manifold Ranking [J].
Guimaraes Pedronette, Daniel Carlos ;
Valem, Lucas Pascotti ;
Almeida, Jurandy ;
Tones, Ricardo da S. .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2019, 28 (12) :5824-5838
[9]   Unsupervised manifold learning through reciprocal kNN graph and Connected Components for image retrieval tasks [J].
Guimaraes Pedronette, Daniel Carlos ;
Fernandes Goncalves, Filipe Marcel ;
Guilherme, Ivan Rizzo .
PATTERN RECOGNITION, 2018, 75 :161-174
[10]  
Han Kai, 2020, P ICLR, P4322