Abstract
Partial label learning (PLL) is a weakly supervised learning method that is able to predict one label as the correct answer from a given candidate label set. In PLL, when all possible candidate labels are as signed to real-world training examples, PLL will hava noisy labeling in its training data set. In the real world, it is unrealistic to assign candidate label to all the training examples. Because semi-supervised partial label learning combines two difficult learning conditions, partial label learning and semi-supervised learning, improving recognition accuracy is a big challenge. Some existing semi-supervised partial label learning boosts the model performance, by assigning to unlabeled data in their label propagation. However, those methods neglect the noisy label in their label propagation, which introduces contaminated data, at the same time it declines model performance. We proposed a semi-supervised partial label learning (SeePLL) method to address the label contamination issue in PLL through reliable label propagation. Specifically, our SeePLL conducts label propagation on the reliable label training set, which filters unreliable data from raw partial label data. SeePLL iteratively updates the unlabeled training set by the reliable label propagation. This iterative manner significantly improves the disambiguation of the unlabeled data. We evaluate the performance of our method on five real-world datasets: Lost, Msrcv2, Mirflickr, BirdSong, and Soccer Player. The experimental results show our method achieves a superior performance than the baselines with a large margin. More importantly, our SeePLL keeps the consistent performance in small proportion of partial label training data resources.
Similar content being viewed by others
References
Chen CH, Patel VM, Chellappa R (2017) Learning from ambiguously labeled face images. IEEE Trans Pattern Anal Mach Intell 40(7):1653–1667
Gong X, Yuan D, Bao W (2021) Top-k partial label machine. IEEE Trans Neural Netw Learn Syst
Lyu G, Feng S, Wang T, et al. (2020) A self-paced regularization framework for partial-label learning. IEEE Trans Cybern Progressiveidentificationoftruelabelsforpartial-labellearning(2020)
LyuG,FengS,LiY,etal.(2020)Hera:partiallabellearningbycombiningheterogeneouslosswith sparseandlow-rankregularization.ACMTransIntellSystTechnol(TIST) 11(3):1–19
LyuG,FengS,HuangW,etal.(2020)Partiallabellearningvialow-rankrepresentationandlabel propagation.SoftComput 24(7):5165–5176
XuS,YangM,ZhouY,etal.(2020)Partiallabelmetriclearningbycollapsingclasses.Int JMachLearnCybern 11:2453–2460
SunK,MinZ,WangJ(2019)PP-PLLprobabilitypropagationforpartiallabellearning.In: Joint europeanconferenceonmachinelearningandknowledgediscoveryindatabases.Springer,Cham,pp123–137
RenX,HeW,QuM,etal.(2016)Afet:Automaticfine-grainedentitytypingbyhierarchical partial-labelembedding.In: Proceedingsofthe2016conferenceonempiricalmethodsinnaturallanguageprocessing, pp 1369–1378
RenX,HeW,QuM,etal.(2016)Labelnoisereductioninentitytypingbyheterogeneouspartial-label embedding.In: Proceedingsofthe22ndACMSIGKDDinternationalconferenceonKnowledgediscoveryanddata mining,pp 1825–1834
ZhouD,ZhangZ,ZhangML,etal.(2018)WeaklysupervisedPOStaggingwithoutdisambiguation. ACMTransAsianLow-ResourLangInfProcess(TALLIP) 17(4):1–19
XieMK,HuangSJ(2021)Partialmulti-labellearningwithnoisylabelidentification.IEEETransPattern AnalMachIntell
XuN,QiaoC,GengX,etal.(2021)Instance-dependentpartiallabellearning.AdvNeuralInfProcess Syst:34
VanEngelenJE,HoosHH(2020)Asurveyonsemi-supervisedlearning.MachLearn 109 (2):373–440
BerthelotD,CarliniN,GoodfellowI,etal.(2019)Mixmatch:aholisticapproachtosemi-supervised learning.arXiv:1905.02249
ZhouZH(2018)Abriefintroductiontoweaklysupervisedlearning.NatlSciRev 5(1):44–53
WangQW,LiYF,ZhouZH(2019)Partiallabellearningwithunlabeleddata.In: IJCAI,pp 3755–3761
JinR,GhahramaniZ(2002)Learningwithmultiplelabels.In: NIPS,vol2.pp 897–904
HüllermeierE,BeringerJ(2006)Learningfromambiguouslylabeledexamples.IntellData Anal 10(5):419–439
WuX,ZhangML(2018)Towardsenablingbinarydecompositionforpartiallabellearning.In: IJCAI, pp 2868–2874
ZhangML,YuF,TangCZ(2017)Disambiguation-freepartiallabellearning.IEEETrans KnowlDataEng 29(10):2155–2167
FengL,AnB(2018)Leveraginglatentlabeldistributionsforpartiallabellearning.In: IJCAI,2107–2113
WangDB,LiL,ZhangML(2019)Adaptivegraphguideddisambiguationforpartiallabellearning.In: Proceedingsofthe25thACMSIGKDDinternationalconferenceonknowledgediscoverydatamining,pp 83–91
ZhangML,ZhouBB,LiuXY(2016)Partiallabellearningviafeature-awaredisambiguation.In: Proceedings ofthe22ndACMSIGKDDinternationalconferenceonknowledgediscoveryanddatamining,pp 1335–1344
MarussyK,BuzaK(2013)Success:anewapproachforsemi-supervisedclassificationoftime-series.In: Internationalconferenceonartificialintelligenceandsoftcomputing.Springer,Berlin,pp437–447
LiYF,LiangDM(2019)Safesemi-supervisedlearning:abriefintroduction.FrontComput Sci 13(4):669–676
BennettK,DemirizA(1999)Semi-supervisedsupportvectormachines.AdvNeuralInfProcessSyst:368–374
LiYF,LiangDM(2019)Safesemi-supervisedlearning:abriefintroduction.FrontComput Sci 13(4):669–676
YuX,HanB,YaoJ,etal.(2019)Howdoesdisagreementhelpgeneralizationagainstlabelcorruption?. In: Internationalconferenceonmachinelearning,PMLR,pp 7164–7173
LiW,MengW,AuMH(2631)Enhancingcollaborativeintrusiondetectionviadisagreement-based semi-supervisedlearninginIoTenvironments.JNetwComputAppl 161(10):2020
ZhangWWML(2020)Semi-supervisedpartiallabellearningviaconfidence-ratedmarginmaximization
ZhangML,YuF(2015)Solvingthepartiallabellearningproblem:aninstance-basedapproach.In: IJCAI, pp 4048–4054
HuiskesMJ,LewMS(2008)Themirflickrretrievalevaluation.In: Proceedingsofthe1stACMinternational conferenceonmultimediainformationretrieval,pp 39–43
BriggsF,FernXZ,RaichR(2012)Rank-losssupportinstancemachinesforMIMLinstanceannotation.In: Proceedingsofthe18thACMSIGKDDinternationalconferenceonknowledgediscoveryanddatamining,pp 534–542
NguyenN,CaruanaR(2008)Classificationwithpartiallabels.In: Proceedingsofthe14thACMSIGKDD internationalconferenceonknowledgediscoveryanddatamining,pp 551–559
Acknowledgments
This work was supported in part by the National Natural Science Foundation of China (Grant No. 61502404), Natural Science Foundation of Fujian Province of China (Grant Nos. 2020J06027, 2019J01851), Distinguished Young Scholars Foundation of Fujian Educational Committee (Grant No. DYS201707), Xiamen Science and Technology Program (Grant No. 3502Z20183059), Open Fund of Engineering Research Center for Software Testing and Evaluation of Fujian Province, and Open Fund of Key Laboratory of Data mining and Intelligent Recommendation, Fujian Province University.
Author information
Authors and Affiliations
Contributions
Ying Ma: Conceptualization, Methodology, Software. Dayuan Chen: Data curation, Software, Writing-Original draft. Tian Wang: Investigation. Guoqi Li: Writing-Reviewing and Editing. Ming Yan: Visualization, Validation.
Corresponding authors
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Ma, Y., Chen, D., Wang, T. et al. Semi-supervised partial label learning algorithm via reliable label propagation. Appl Intell 53, 12859–12872 (2023). https://doi.org/10.1007/s10489-022-04027-9
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-022-04027-9