Presentation 2013-12-21
Dynamic Binary Neural Networks for Deep learning based on correlation
Jungo MORIYASU, Toshimichi SAITO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper studies basic dynamics and learning capability of the multi-layer dynamic binary neural network. The network has the signum activation function and can exhibit various binary periodic orbits. The network dynamics can be visualized by the Gray-code-based return map. We present a deep learning algorithm based on the correlation learning. The purpose of the learning is not only storage of teacher signal but also enlargement of the domain of attraction to the teacher signal. As a typical example of the teacher signal, we use a periodic orbit corresponding to the control signal of the matrix converters. Performing basic numerical experiment, we have confirmed that the teacher signal can be stored successfully and the deep structure can be effective to enlarge the domain of attraction.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Neural Network / Correlation learning / Genetic Algorithm / Return Map / Deep learning
Paper # NC2013-62
Date of Issue

Conference Information
Committee NC
Conference Date 2013/12/14(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) Dynamic Binary Neural Networks for Deep learning based on correlation
Sub Title (in English)
Keyword(1) Neural Network
Keyword(2) Correlation learning
Keyword(3) Genetic Algorithm
Keyword(4) Return Map
Keyword(5) Deep learning
1st Author's Name Jungo MORIYASU
1st Author's Affiliation EE Dept., Hosei University()
2nd Author's Name Toshimichi SAITO
2nd Author's Affiliation EE Dept., Hosei University
Date 2013-12-21
Paper # NC2013-62
Volume (vol) vol.113
Number (no) 374
Page pp.pp.-
#Pages 5
Date of Issue