Presentation 2005-09-21
Fast Semi-Supervised Learning on Nearest Neighbor Graphs
Weiwei DU, Kiichi URAHAMA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A simple label propagation technique is presented for semi-supervised learning of pattarn classifiers. In our technique, label is propagated on a sparse unweithed graph called the kNN graph. This propagation can be executed with only integer computation and stops in finite steps. Our method is therefore faster than the conventional label propagation methods dealing with real numbers and with slow convergence. We develop an unsupervised learning method exploiting this label propagation procedure and apply it to select data to be labeled in our semi-supervised learning. The performance of the present method is examined for a synthetic dataset and some real datasets.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) semi-supervised learning / label propagation / unsupervised learning / outlier detection / nearest neighbor graph
Paper # NLC2005-24,PRMU2005-51
Date of Issue

Conference Information
Committee NLC
Conference Date 2005/9/14(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Vice Chair

Paper Information
Registration To Natural Language Understanding and Models of Communication (NLC)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Fast Semi-Supervised Learning on Nearest Neighbor Graphs
Sub Title (in English)
Keyword(1) semi-supervised learning
Keyword(2) label propagation
Keyword(3) unsupervised learning
Keyword(4) outlier detection
Keyword(5) nearest neighbor graph
1st Author's Name Weiwei DU
1st Author's Affiliation Faculty of Design, Kyushu University()
2nd Author's Name Kiichi URAHAMA
2nd Author's Affiliation Faculty of Design, Kyushu University
Date 2005-09-21
Paper # NLC2005-24,PRMU2005-51
Volume (vol) vol.105
Number (no) 299
Page pp.pp.-
#Pages 6
Date of Issue