Presentation 2015-01-29
Implementation of the higher-rank of SOM using formal neurons
Yuta SAHO, Kiyohisa NATSUME, Tetsuo FURUKAWA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The purpose of this work is building a neural network model which performs meta-model learning. The term 'meta-model learning' used here means a kind of meta-learning or multi-task learning scheme, aiming discovering general knowledge from individual experiences. For this purpose, the meta-learner needs to learn a set of models organized by the primary learners. It means that the model information of the primary learners should be transferred to the meta-learner. The central issue of this work is what kind of neural circuit is required to transfer the model information, which seems to be stored as synaptic weights in the neurons. In this paper, we propose an implementation using a neural field in which propagation wave traverses. We implemented the higher-rank of self-organizing map by using this model, as a representative algorithm of meta-model learning.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) meta-model learning / meta-learning / multi-task learning / transfer learning / general knowledge discovering / formal neuron / higher-rank of SOM
Paper # NC2014-62
Date of Issue

Conference Information
Committee NC
Conference Date 2015/1/22(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) Implementation of the higher-rank of SOM using formal neurons
Sub Title (in English)
Keyword(1) meta-model learning
Keyword(2) meta-learning
Keyword(3) multi-task learning
Keyword(4) transfer learning
Keyword(5) general knowledge discovering
Keyword(6) formal neuron
Keyword(7) higher-rank of SOM
1st Author's Name Yuta SAHO
1st Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology()
2nd Author's Name Kiyohisa NATSUME
2nd Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology
3rd Author's Name Tetsuo FURUKAWA
3rd Author's Affiliation Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology
Date 2015-01-29
Paper # NC2014-62
Volume (vol) vol.114
Number (no) 437
Page pp.pp.-
#Pages 6
Date of Issue