Recurrent neural network with backpropagation through time for speech recognition
Speech recognition and understanding have been studied for many years. The neural network is well-known as a technique that is able to classify nonlinear problems. Much research has been done in applying neural networks to solving the problem of recognizing speech such as Arabic. Arabic offers a num...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | en |
| Published: |
2004
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/2016/1/paper122.pdf http://eprints.utm.my/2016/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1845470594607349760 |
|---|---|
| author | Ahmad, A. M. Ismail, S. Samaon, D. F. |
| author_facet | Ahmad, A. M. Ismail, S. Samaon, D. F. |
| author_sort | Ahmad, A. M. |
| building | UTM Library |
| collection | Institutional Repository |
| content_provider | Universiti Teknologi Malaysia |
| content_source | UTM Institutional Repository |
| continent | Asia |
| country | Malaysia |
| description | Speech recognition and understanding have been studied for many years. The neural network is well-known as a technique that is able to classify nonlinear problems. Much research has been done in applying neural networks to solving the problem of recognizing speech such as Arabic. Arabic offers a number of challenges to speech recognition. We propose a fully-connected hidden layer between the input and state nodes and the output. We also investigate and show that this hidden layer makes the learning of complex classification tasks more efficient. We also investigate the difference between LPCC (linear predictive cepstrum coefficients) and MFCC (Mel-frequency cepstral coefficients) in the feature extraction process. The aim of the study was to observe the differences in the 29 letters of the Arabic alphabet from "alif" to "ya". The purpose of this research is to upgrade the knowledge and understanding of Arabic alphabet or words using a fully-connected recurrent neural network (FCRNN) and backpropagation through time (BPTT) learning algorithm. Six speakers (a mixture of male and female) in a quiet environment are used in training. |
| format | Conference or Workshop Item |
| id | my.utm.eprints-2016 |
| institution | Universiti Teknologi Malaysia |
| language | en |
| publishDate | 2004 |
| record_format | eprints |
| spelling | my.utm.eprints-20162017-09-10T08:21:37Z http://eprints.utm.my/2016/ Recurrent neural network with backpropagation through time for speech recognition Ahmad, A. M. Ismail, S. Samaon, D. F. TK Electrical engineering. Electronics Nuclear engineering Speech recognition and understanding have been studied for many years. The neural network is well-known as a technique that is able to classify nonlinear problems. Much research has been done in applying neural networks to solving the problem of recognizing speech such as Arabic. Arabic offers a number of challenges to speech recognition. We propose a fully-connected hidden layer between the input and state nodes and the output. We also investigate and show that this hidden layer makes the learning of complex classification tasks more efficient. We also investigate the difference between LPCC (linear predictive cepstrum coefficients) and MFCC (Mel-frequency cepstral coefficients) in the feature extraction process. The aim of the study was to observe the differences in the 29 letters of the Arabic alphabet from "alif" to "ya". The purpose of this research is to upgrade the knowledge and understanding of Arabic alphabet or words using a fully-connected recurrent neural network (FCRNN) and backpropagation through time (BPTT) learning algorithm. Six speakers (a mixture of male and female) in a quiet environment are used in training. 2004 Conference or Workshop Item PeerReviewed application/pdf en http://eprints.utm.my/2016/1/paper122.pdf Ahmad, A. M. and Ismail, S. and Samaon, D. F. (2004) Recurrent neural network with backpropagation through time for speech recognition. In: IEEE International Symposium on Communications and Information Technology, 2004. ISCIT 2004. . |
| spellingShingle | TK Electrical engineering. Electronics Nuclear engineering Ahmad, A. M. Ismail, S. Samaon, D. F. Recurrent neural network with backpropagation through time for speech recognition |
| title | Recurrent neural network with backpropagation through time for speech recognition |
| title_full | Recurrent neural network with backpropagation through time for speech recognition |
| title_fullStr | Recurrent neural network with backpropagation through time for speech recognition |
| title_full_unstemmed | Recurrent neural network with backpropagation through time for speech recognition |
| title_short | Recurrent neural network with backpropagation through time for speech recognition |
| title_sort | recurrent neural network with backpropagation through time for speech recognition |
| topic | TK Electrical engineering. Electronics Nuclear engineering |
| url | http://eprints.utm.my/2016/1/paper122.pdf http://eprints.utm.my/2016/ |
| url_provider | http://eprints.utm.my/ |