Supervised Transfer Sparse Coding

Handle URI:
http://hdl.handle.net/10754/556535
Title:
Supervised Transfer Sparse Coding
Authors:
Al-Shedivat, Maruan ( 0000-0001-9037-1005 ) ; Wang, Jim Jing-Yan; Alzahrani, Majed; Huang, Jianhua Z.; Gao, Xin ( 0000-0002-7108-3574 )
Abstract:
A combination of the sparse coding and transfer learn- ing techniques was shown to be accurate and robust in classification tasks where training and testing objects have a shared feature space but are sampled from differ- ent underlying distributions, i.e., belong to different do- mains. The key assumption in such case is that in spite of the domain disparity, samples from different domains share some common hidden factors. Previous methods often assumed that all the objects in the target domain are unlabeled, and thus the training set solely comprised objects from the source domain. However, in real world applications, the target domain often has some labeled objects, or one can always manually label a small num- ber of them. In this paper, we explore such possibil- ity and show how a small number of labeled data in the target domain can significantly leverage classifica- tion accuracy of the state-of-the-art transfer sparse cod- ing methods. We further propose a unified framework named supervised transfer sparse coding (STSC) which simultaneously optimizes sparse representation, domain transfer and classification. Experimental results on three applications demonstrate that a little manual labeling and then learning the model in a supervised fashion can significantly improve classification accuracy.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Publisher:
The AAAI Press
Journal:
Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligence
Conference/Event name:
the Twenty-Eighth AAAI Conference on Artificial Intelligence, 2014
Issue Date:
27-Jul-2014
Type:
Conference Paper
Sponsors:
the Association for the Advancement of Artificial Intelligence
Additional Links:
http://maruan.alshedivat.com/wp-content/uploads/2014/07/STSC.pdf
Appears in Collections:
Conference Papers; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorAl-Shedivat, Maruanen
dc.contributor.authorWang, Jim Jing-Yanen
dc.contributor.authorAlzahrani, Majeden
dc.contributor.authorHuang, Jianhua Z.en
dc.contributor.authorGao, Xinen
dc.date.accessioned2015-06-07T21:45:04Zen
dc.date.available2015-06-07T21:45:04Zen
dc.date.issued2014-07-27en
dc.identifier.urihttp://hdl.handle.net/10754/556535en
dc.description.abstractA combination of the sparse coding and transfer learn- ing techniques was shown to be accurate and robust in classification tasks where training and testing objects have a shared feature space but are sampled from differ- ent underlying distributions, i.e., belong to different do- mains. The key assumption in such case is that in spite of the domain disparity, samples from different domains share some common hidden factors. Previous methods often assumed that all the objects in the target domain are unlabeled, and thus the training set solely comprised objects from the source domain. However, in real world applications, the target domain often has some labeled objects, or one can always manually label a small num- ber of them. In this paper, we explore such possibil- ity and show how a small number of labeled data in the target domain can significantly leverage classifica- tion accuracy of the state-of-the-art transfer sparse cod- ing methods. We further propose a unified framework named supervised transfer sparse coding (STSC) which simultaneously optimizes sparse representation, domain transfer and classification. Experimental results on three applications demonstrate that a little manual labeling and then learning the model in a supervised fashion can significantly improve classification accuracy.en
dc.description.sponsorshipthe Association for the Advancement of Artificial Intelligenceen
dc.publisherThe AAAI Pressen
dc.relation.urlhttp://maruan.alshedivat.com/wp-content/uploads/2014/07/STSC.pdfen
dc.rightsArchived with thanks to Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligenceen
dc.titleSupervised Transfer Sparse Codingen
dc.typeConference Paperen
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.identifier.journalProceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligenceen
dc.conference.dateJuly 27 –31, 2014en
dc.conference.namethe Twenty-Eighth AAAI Conference on Artificial Intelligence, 2014en
dc.conference.locationQuébec City, Québec, Canadaen
dc.eprint.versionPost-printen
dc.contributor.institutionUniversity at Buffalo, The State University of New York, Buffalo, NY 14203, United Statesen
dc.contributor.institutionDepartment of Statistics, Texas A&M University, College Station, TX 77843, United Statesen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.