Title :
Recognition of Korean spoken digit using single layer recurrent neural networks
Author :
Ryeu, J.K. ; Tak, H.Y. ; Heo, N.W. ; Chung, H.S.
Author_Institution :
Dept. of Electr. Eng., Kyungpook Nat. Univ., Taegu, South Korea
Abstract :
We proposed the novel single layer dynamic and static recurrent neural networks. These networks can be used to the recognition of dynamic patterns. With the feedback connections, the learning performance of the patterns can be improved. Considering the hardware implementation, we used integer type learning algorithm. These networks were applied to the recognition of Korean spoken digits.
Keywords :
feedback; learning (artificial intelligence); recurrent neural nets; speech recognition; Korean spoken digit recognition; dynamic recurrent neural networks; feedback connections; integer type learning algorithm; learning performance; single-layer recurrent neural networks; static recurrent neural networks; Hardware; Marketing and sales; Multi-layer neural network; Network address translation; Neural networks; Neurofeedback; Pattern recognition; Recurrent neural networks; Speech recognition; Testing;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.713908
