CASIA OpenIR  > 模式识别国家重点实验室  > 图像与视频分析
BundleNet Learning with Noisy Label via Sample Correlations
Li, Chenghua1,4; Zhang, Chunjie2,5; Ding, Kun2,4; Li, Gang1,4; Cheng, Jian1,4,6; Lu, Hanqing3,4; Jian Cheng
Source PublicationIEEE ACCESS
AbstractSequential patterns are important, because they can be exploited to improve the prediction accuracy of our classifiers. Sequential data, such as time series/video frames, and event data are becoming more and more ubiquitous in a wide spectrum of application scenarios especially in the background of large data and deep learning. However, large data sets used in training modern machine-learning models, such as deep neural networks, are often affected by label noise. Existing noisy learning approaches mainly focus on building an additional network to clean the noise or find a robust loss function. Few works tackle this problem by exploiting sample correlations. In this paper, we propose BundleNet, a framework of sequential structure (named bundle-module, see Fig. 1) for deep neural networks to handle the label noise. The bundle module naturally takes into account sample correlations by constructing bundles of samples class-by-class, and treats them as independent inputs. Moreover, we prove that the bundle-module performs a form of regularization, which is similar to dropout as regularization during training. The regularization effect endows the BundleNet with strong robustness to the label noise. Extensive experiments on public data sets prove that the proposed approach is effective and promising.
KeywordBundlenet Sequential Data Classification Noisy Label Regularization
WOS HeadingsScience & Technology ; Technology
Indexed BySCI
Funding OrganizationNational Natural Science Foundation of China(61332016) ; Jiangsu Key Laboratory of Big Data Analysis Technology ; 863 Program(2014AA015105)
WOS Research AreaComputer Science ; Engineering ; Telecommunications
WOS SubjectComputer Science, Information Systems ; Engineering, Electrical & Electronic ; Telecommunications
WOS IDWOS:000426275700001
Citation statistics
Cited Times:1[WOS]   [WOS Record]     [Related Records in WOS]
Document Type期刊论文
Corresponding AuthorJian Cheng
Affiliation1.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China
2.Chinese Acad Sci, Inst Automat, Pattern Recognit & Intelligent Syst, Beijing 100190, Peoples R China
3.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
4.Univ Chinese Acad Sci, Beijing 100049, Peoples R China
5.Univ Chinese Acad Sci, Sch Comp & Control Engn, Beijing 100049, Peoples R China
6.Chinese Acad Sci, Inst Automat, Res Ctr Brain Inspired Intelligence, Beijing 100190, Peoples R China
First Author AffilicationInstitute of Automation, Chinese Academy of Sciences
Recommended Citation
GB/T 7714
Li, Chenghua,Zhang, Chunjie,Ding, Kun,et al. BundleNet Learning with Noisy Label via Sample Correlations[J]. IEEE ACCESS,2018,6(1):2367-2377.
APA Li, Chenghua.,Zhang, Chunjie.,Ding, Kun.,Li, Gang.,Cheng, Jian.,...&Jian Cheng.(2018).BundleNet Learning with Noisy Label via Sample Correlations.IEEE ACCESS,6(1),2367-2377.
MLA Li, Chenghua,et al."BundleNet Learning with Noisy Label via Sample Correlations".IEEE ACCESS 6.1(2018):2367-2377.
Files in This Item: Download All
File Name/Size DocType Version Access License
BundleNet.pdf(1606KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Li, Chenghua]'s Articles
[Zhang, Chunjie]'s Articles
[Ding, Kun]'s Articles
Baidu academic
Similar articles in Baidu academic
[Li, Chenghua]'s Articles
[Zhang, Chunjie]'s Articles
[Ding, Kun]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Li, Chenghua]'s Articles
[Zhang, Chunjie]'s Articles
[Ding, Kun]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: BundleNet.pdf
Format: Adobe PDF
This file does not support browsing at this time
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.