Presentation 2000/10/13
An LSI Fabrication of Quantized Connection Neural Networks and Its Learning
Y. Katayama, S. Sato, K. Nakajima,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a large fan-in majority circuit. The proposed circuit configulation is so simple that is implemented by standard CMOS technology. This circuit comprises N input parallel stages (N is the number of inputs and is able to be increased over 1000), an output-buffer and a reference voltage generator. An important application of the majority circuits is a formation of binary neural networks. It needs the huge number of neurons. Thefunctions of a binary neuron and a majority circuit are essentially the same. Therefore, the proposed circuit makes it possible to realize hardware implementation of such neural networks.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Neural Networks / LSI / Quantized Connection / DBM Learning
Paper # NLP2000-73
Date of Issue

Conference Information
Committee NLP
Conference Date 2000/10/13(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 Nonlinear Problems (NLP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An LSI Fabrication of Quantized Connection Neural Networks and Its Learning
Sub Title (in English)
Keyword(1) Neural Networks
Keyword(2) LSI
Keyword(3) Quantized Connection
Keyword(4) DBM Learning
1st Author's Name Y. Katayama
1st Author's Affiliation Laboratory for Electronic Intelligent Systems, R.I.E.C., Tohoku University()
2nd Author's Name S. Sato
2nd Author's Affiliation Laboratory for Electronic Intelligent Systems, R.I.E.C., Tohoku University
3rd Author's Name K. Nakajima
3rd Author's Affiliation Laboratory for Electronic Intelligent Systems, R.I.E.C., Tohoku University
Date 2000/10/13
Paper # NLP2000-73
Volume (vol) vol.100
Number (no) 381
Page pp.pp.-
#Pages 8
Date of Issue