Presentation 2002/11/6
The pulse mode neural network using the ramp function neuron with additive noise
Shinsuke KURODA, Hiroomi HIKAWA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper proposes a new type of digital pulse mode neuron with smooth nonlinear activation function, which is obtained by adding noise sequence to ramp function neuron. To verify the feasibility of the proposed neuron, various computer simulations are carried out. The multilayer neural network with on-chip learning is designed with the proposed neuron. On-chip learning circuit is improved to have better learning capability. Using exclusive OR and 2-D binary classifying problem, the effect of the neuron and learning circuit on the learning capability is studied by computer simulation. The results show that the proposed neuron and learning circuit can improve the learning capability.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Ramp function / Pulse mode neural network / Back-propagation / Additive noise
Paper # NLP2002-81
Date of Issue

Conference Information
Committee NLP
Conference Date 2002/11/6(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) The pulse mode neural network using the ramp function neuron with additive noise
Sub Title (in English)
Keyword(1) Ramp function
Keyword(2) Pulse mode neural network
Keyword(3) Back-propagation
Keyword(4) Additive noise
1st Author's Name Shinsuke KURODA
1st Author's Affiliation Dept of Computer Science and Intelligent Systems, Oita University()
2nd Author's Name Hiroomi HIKAWA
2nd Author's Affiliation Dept of Computer Science and Intelligent Systems, Oita University
Date 2002/11/6
Paper # NLP2002-81
Volume (vol) vol.102
Number (no) 432
Page pp.pp.-
#Pages 6
Date of Issue