A Novel Transfer Learning Method Based on Common Space Mapping and Weighted Domain Matching

Handle URI:
http://hdl.handle.net/10754/622934
Title:
A Novel Transfer Learning Method Based on Common Space Mapping and Weighted Domain Matching
Authors:
Liang, Ru-Ze; Xie, Wei; Li, Weizhi; Wang, Hongqi; Wang, Jim Jing-Yan; Taylor, Lisa
Abstract:
In this paper, we propose a novel learning framework for the problem of domain transfer learning. We map the data of two domains to one single common space, and learn a classifier in this common space. Then we adapt the common classifier to the two domains by adding two adaptive functions to it respectively. In the common space, the target domain data points are weighted and matched to the target domain in term of distributions. The weighting terms of source domain data points and the target domain classification responses are also regularized by the local reconstruction coefficients. The novel transfer learning framework is evaluated over some benchmark cross-domain data sets, and it outperforms the existing state-of-the-art transfer learning methods.
KAUST Department:
King Abdullah University of Science and Technology, Saudi Arabia
Citation:
Liang R-Z, Xie W, Li W, Wang H, Wang JJ-Y, et al. (2016) A Novel Transfer Learning Method Based on Common Space Mapping and Weighted Domain Matching. 2016 IEEE 28th International Conference on Tools with Artificial Intelligence (ICTAI). Available: http://dx.doi.org/10.1109/ICTAI.2016.0053.
Publisher:
Institute of Electrical and Electronics Engineers (IEEE)
Journal:
2016 IEEE 28th International Conference on Tools with Artificial Intelligence (ICTAI)
Issue Date:
17-Jan-2017
DOI:
10.1109/ICTAI.2016.0053
Type:
Conference Paper
Additional Links:
http://ieeexplore.ieee.org/document/7814613/
Appears in Collections:
Conference Papers

Full metadata record

DC FieldValue Language
dc.contributor.authorLiang, Ru-Zeen
dc.contributor.authorXie, Weien
dc.contributor.authorLi, Weizhien
dc.contributor.authorWang, Hongqien
dc.contributor.authorWang, Jim Jing-Yanen
dc.contributor.authorTaylor, Lisaen
dc.date.accessioned2017-02-28T11:54:07Z-
dc.date.available2017-02-28T11:54:07Z-
dc.date.issued2017-01-17en
dc.identifier.citationLiang R-Z, Xie W, Li W, Wang H, Wang JJ-Y, et al. (2016) A Novel Transfer Learning Method Based on Common Space Mapping and Weighted Domain Matching. 2016 IEEE 28th International Conference on Tools with Artificial Intelligence (ICTAI). Available: http://dx.doi.org/10.1109/ICTAI.2016.0053.en
dc.identifier.doi10.1109/ICTAI.2016.0053en
dc.identifier.urihttp://hdl.handle.net/10754/622934-
dc.description.abstractIn this paper, we propose a novel learning framework for the problem of domain transfer learning. We map the data of two domains to one single common space, and learn a classifier in this common space. Then we adapt the common classifier to the two domains by adding two adaptive functions to it respectively. In the common space, the target domain data points are weighted and matched to the target domain in term of distributions. The weighting terms of source domain data points and the target domain classification responses are also regularized by the local reconstruction coefficients. The novel transfer learning framework is evaluated over some benchmark cross-domain data sets, and it outperforms the existing state-of-the-art transfer learning methods.en
dc.publisherInstitute of Electrical and Electronics Engineers (IEEE)en
dc.relation.urlhttp://ieeexplore.ieee.org/document/7814613/en
dc.rights(c) 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.en
dc.subjectlearning (artificial intelligence)en
dc.subjectpattern classificationen
dc.subjectBenchmark testingen
dc.subjectDVDen
dc.subjectElectronic mailen
dc.subjectLearning systemsen
dc.subjectMinimizationen
dc.subjectOptimizationen
dc.subjectTrainingen
dc.titleA Novel Transfer Learning Method Based on Common Space Mapping and Weighted Domain Matchingen
dc.typeConference Paperen
dc.contributor.departmentKing Abdullah University of Science and Technology, Saudi Arabiaen
dc.identifier.journal2016 IEEE 28th International Conference on Tools with Artificial Intelligence (ICTAI)en
dc.eprint.versionPost-printen
dc.contributor.institutionVanderbilt University, Nashville, TN 37235, United Statesen
dc.contributor.institutionSuning Commerce R&D Center USA, Inc Palo Alto, CA 94304, United Statesen
dc.contributor.institutionSchool of Management of Harbin University of Science and Technology, Harbin 150000, Chinaen
dc.contributor.institutionNew York University Abu Dhabi, United Arab Emiratesen
dc.contributor.institutionMichigan State University, East Lansing, MI 48824, United Statesen
kaust.authorLiang, Ru-Zeen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.