Presentation 2009-01-19
Node perturbation learning with noisy reference
Tatsuya CHO, Kentaro KATAHIRA, Masato OKADA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a node perturbation learning with noisy reference signal. Recently, the method for node perturbation has investigated without consideration of noise in reference signal. In biological system, however, noise plays an essential role and neural activities are intrinsically noisy. In this paper, we analyze a node perturbation with noisy reference used with linear perceptron. Learning succeeds even with the noisy reference. Proposed learning scheme reduces residual error in factor of output units compared to the noiseless cases.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) noisy reference / node perturbation / stocastic gradient method / liner perceptron
Paper # NC2008-89
Date of Issue

Conference Information
Committee NC
Conference Date 2009/1/12(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Node perturbation learning with noisy reference
Sub Title (in English)
Keyword(1) noisy reference
Keyword(2) node perturbation
Keyword(3) stocastic gradient method
Keyword(4) liner perceptron
1st Author's Name Tatsuya CHO
1st Author's Affiliation Graduate School of Frontier Sciences, The University of Tokyo()
2nd Author's Name Kentaro KATAHIRA
2nd Author's Affiliation Graduate School of Frontier Sciences, The University of Tokyo:RIKEN Brain Science Institute
3rd Author's Name Masato OKADA
3rd Author's Affiliation Graduate School of Frontier Sciences, The University of Tokyo:RIKEN Brain Science Institute
Date 2009-01-19
Paper # NC2008-89
Volume (vol) vol.108
Number (no) 383
Page pp.pp.-
#Pages 5
Date of Issue