Abstract | ||
---|---|---|
Under the guide of the novel biomimetics pattern recognition theory that is based on the character of human's recognition, combining the character of traditional neural network, a new multi-weight neural network is constructed to realize the idea of that theory. Extraordinary results are obtained with the first use of the new multi-weight neural network in speech recognition of finite words. The experiment results show that the multi-weight neural network can not only recognize finite words correctly and promptly, but also keep quite high correct recognition rate under the circumstance of small-number samples. |
Year | DOI | Venue |
---|---|---|
2005 | 10.1007/11427445_55 | ISNN (2) |
Keywords | Field | DocType |
small-number sample,speech recognition,multi-weight neural network,finite word,new multi-weight neural network,experiment result,novel biomimetics pattern recognition,extraordinary result,high correct recognition rate,traditional neural network,neural network,pattern recognition | Neocognitron,Feature vector,Radial basis function,Pattern recognition,Intelligent character recognition,Computer science,Recurrent neural network,Speech recognition,Time delay neural network,Artificial intelligence,Artificial neural network,Neural gas | Conference |
Volume | ISSN | ISBN |
3497 | 0302-9743 | 3-540-25913-9 |
Citations | PageRank | References |
0 | 0.34 | 1 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yan Wu | 1 | 20 | 4.50 |
Hongbo Wang | 2 | 0 | 0.34 |
Mingxi Jin | 3 | 0 | 0.68 |
Shoujue Wang | 4 | 86 | 25.23 |