A filter neural network

Author

Yong, Lim Kia ; En, Cao ; Zhou Rajing ; Aun, Ng Kien

Author_Institution

Nanyang Technol. Inst., Singapore

Volume

1

fYear

1993

fDate

25-29 Oct. 1993

Firstpage

371

Abstract

This paper proposes to add a filter layer to a dot product matching neural network. The purpose of the filter layer is to discard those unfavourable choices by checking the lower and upper bounds of each exemplar with the test pattern. The product is a supervised, fast learning filter neural network. It has a better generalisation capability than an ordinary dot product matching neural network. The new neural network is tested for speaker-independent spoken number (in English) recognition. An accuracy of 96.5% is reported for the test data. Without the filter layer, the recognition rate falls to 94.0%.

Keywords

feedforward neural nets; filtering theory; generalisation (artificial intelligence); learning (artificial intelligence); pattern classification; speech recognition; dot product matching neural network; filter neural network; generalisation capability; recognition rate; speaker-independent spoken number recognition; supervised fast learning filter neural network; test pattern; Energy states; Feedforward neural networks; Feeds; Filtering; Matched filters; Neural networks; Pattern matching; Pattern recognition; Testing; Upper bound;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on

Print_ISBN

0-7803-1421-2

Type

conf

DOI

10.1109/IJCNN.1993.713934

Filename

713934