Presentation | 2005-05-13 Implementation of Simplified Back Propagation Algorithm on a Dynamically Reconfigurable Device PCA-2 Tomoko OOTSUKA, Keigo KURATA, Junji KITAMICHI, Kenichi KURODA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | We have been studying the implementation method of neural networks on Plastic Cell Architecture (PCA), a fine-grained dynamically reconfigurable device, to execute neural networks processing effectively. Conventional PCA neuron models, which we proposed for multi-layer neural networks, execute only static processing. In this paper, we propose a implementation model of multi-layer neural networks with learning dynamically based on the back propagation algorithm. A proposed model can execute a learning process in parallel with a forward process. Connection information between neurons that each neuron must hold are reduced to simplify the learning process. Actually, we designed the proposed model on PCA-2, and simulated parallel processing of the learning process and the forward process. The proposed model uses dynamic routing functions of PCA with small routing area. Additionally, we discussed issues revealed by evaluation. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Dynamically Reconfigurable Device / Neural Network / Back Propagation Algorithm |
Paper # | RECONF2005-27 |
Date of Issue |
Conference Information | |
Committee | RECONF |
---|---|
Conference Date | 2005/5/6(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 | Reconfigurable Systems (RECONF) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Implementation of Simplified Back Propagation Algorithm on a Dynamically Reconfigurable Device PCA-2 |
Sub Title (in English) | |
Keyword(1) | Dynamically Reconfigurable Device |
Keyword(2) | Neural Network |
Keyword(3) | Back Propagation Algorithm |
1st Author's Name | Tomoko OOTSUKA |
1st Author's Affiliation | Graduate School of Computer Science and Engineering, The University of Aizu() |
2nd Author's Name | Keigo KURATA |
2nd Author's Affiliation | Graduate School of Computer Science and Engineering, The University of Aizu |
3rd Author's Name | Junji KITAMICHI |
3rd Author's Affiliation | Graduate School of Computer Science and Engineering, The University of Aizu |
4th Author's Name | Kenichi KURODA |
4th Author's Affiliation | Graduate School of Computer Science and Engineering, The University of Aizu |
Date | 2005-05-13 |
Paper # | RECONF2005-27 |
Volume (vol) | vol.105 |
Number (no) | 43 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |