DocumentCode :
2264055
Title :
Error-correcting semi-supervised learning with mode-filter on graphs
Author :
Du, Weiwei ; Urahama, Kiichi
Author_Institution :
Kyoto Inst. of Technol., Kyoto, Japan
fYear :
2009
fDate :
Sept. 27 2009-Oct. 4 2009
Firstpage :
2095
Lastpage :
2100
Abstract :
We present a semi-supervised learning algorithm robust to label errors in training data. Our method employs the mode filter used for smoothing noisy images. We extend it from images to functions on graphs for regression of classification functions on an undirected graph. Our contribution in this paper lies in the introduction of nonlinearity in the regression in contrast to linear interpolation used in previous graph-based semi-supervised learning algorithms. Error-correcting effect of mode filters is demonstrated and the classification rates of the present learning method is evaluated with experiments for the UCI benchmark datasets contaminated with label errors.
Keywords :
graphs; image denoising; interpolation; learning (artificial intelligence); UCI benchmark datasets; classification functions; error-correcting semi-supervised learning; graph-based semi-supervised learning; linear interpolation; mode-filter; noisy image smoothing; undirected graph; Conferences; Filters; Gray-scale; Iterative algorithms; Iterative methods; Laplace equations; Noise robustness; Semisupervised learning; Smoothing methods; Training data;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computer Vision Workshops (ICCV Workshops), 2009 IEEE 12th International Conference on
Conference_Location :
Kyoto
Print_ISBN :
978-1-4244-4442-7
Electronic_ISBN :
978-1-4244-4441-0
Type :
conf
DOI :
10.1109/ICCVW.2009.5457539
Filename :
5457539
Link To Document :
بازگشت