Presentation | 2000/6/15 Disambiguation of Part-of-Speech with Modular Neural Networks Qing MA, Bao-Liang LU, Hitoshi ISAHARA, Michinori ICHIKAWA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | A part of speech tagging system using neural networks has been proposed by Ma et al. The system can do part of speech tagging with accuracy far higher than that of using Hidden Markov Model, a main approach for part of speech tagging, when a small Thai corpus with ten thousand order words was used for learning. Because of the convergence problem existing in multilayer perceptron used, however, the system was slow to converge and took a long time to learn even the above mentioned small amount of learning data. It would therefore be difficult to further improve its accuracy by increasing the number of times of learning or increasing the number of learning data. To resolve this problem, this paper adopted the Min-Max Modular Neural Networks proposed by Lu et al.in the part of speech tagging system. By using this modular neural network, which can break down the part of speech tagging problems into a number of independent smaller and simpler subproblems and all of the subproblems can be learned by small network modules in parallel, there will be no convergence problem and the part of speech tagging problems can be learned more fast and more accurate. As a result, it turned possible to increase the learning accuracy using the same learning data and to learn larger amount of data, and it therefore turned possible to further improve the tagging accuracy. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | part of speech tagging / Thai corpus / modular neural networks / parallel learning / over leaning |
Paper # | NLP2000-23,NC2000-17 |
Date of Issue |
Conference Information | |
Committee | NLP |
---|---|
Conference Date | 2000/6/15(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) | Disambiguation of Part-of-Speech with Modular Neural Networks |
Sub Title (in English) | |
Keyword(1) | part of speech tagging |
Keyword(2) | Thai corpus |
Keyword(3) | modular neural networks |
Keyword(4) | parallel learning |
Keyword(5) | over leaning |
1st Author's Name | Qing MA |
1st Author's Affiliation | Kansai Advanced Research Center, Communications Research Laboratory() |
2nd Author's Name | Bao-Liang LU |
2nd Author's Affiliation | RIKEN Brain Science Institute |
3rd Author's Name | Hitoshi ISAHARA |
3rd Author's Affiliation | Kansai Advanced Research Center, Communications Research Laboratory |
4th Author's Name | Michinori ICHIKAWA |
4th Author's Affiliation | RIKEN Brain Science Institute |
Date | 2000/6/15 |
Paper # | NLP2000-23,NC2000-17 |
Volume (vol) | vol.100 |
Number (no) | 124 |
Page | pp.pp.- |
#Pages | 8 |
Date of Issue |