Presentation | 1996/3/14 Avoiding paralysis in the backpropagation algorithm Yadira Solano, Hiroaki IKEDA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | The backpropagation algorithm is now the most widely used algorithm in artificial neural networks. However, it is not without its real problems, the most serious being the slow speed of learning which sometimes produces a paralysis of the network, as learning virtually stops. In the present article a modification to the backpropagation algorithm for learning in neural networks is proposed. The results obtained from the evaluation of the performance of both algo-rithms: the original and the one modified in our proposal, are discussed. It is shown that our proposal not only reduces significant dependency on the initial parameters but contributes to speed up convergence to a given error value. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | neural networks / backpropagation / learning / paralysis |
Paper # | AI95-63,KBSE95-51 |
Date of Issue |
Conference Information | |
Committee | AI |
---|---|
Conference Date | 1996/3/14(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 | Artificial Intelligence and Knowledge-Based Processing (AI) |
---|---|
Language | ENG |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Avoiding paralysis in the backpropagation algorithm |
Sub Title (in English) | |
Keyword(1) | neural networks |
Keyword(2) | backpropagation |
Keyword(3) | learning |
Keyword(4) | paralysis |
1st Author's Name | Yadira Solano |
1st Author's Affiliation | Faculty of Natural Sciences, Graduate School of Science and Technology, Chiba University() |
2nd Author's Name | Hiroaki IKEDA |
2nd Author's Affiliation | Faculty of Natural Sciences, Graduate School of Science and Technology, Chiba University |
Date | 1996/3/14 |
Paper # | AI95-63,KBSE95-51 |
Volume (vol) | vol.95 |
Number (no) | 573 |
Page | pp.pp.- |
#Pages | 8 |
Date of Issue |