A Hybrid Architecture for Accent-Based Automatic Speech Recognition Systems for E-Learning Environment
The adoption of accents-based speech recognition into the e- learning environment has revolutionized the e-learning technology and reducing the learners’ barriers to knowledge acquisition, particularly as it relates to the effects of accents on the participants’ comprehension. Several researchers ha...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Stefan cel Mare University of Suceava
2022-10-01
|
| Series: | Journal of Applied Computer Science & Mathematics |
| Subjects: | |
| Online Access: | https://jacsm.ro/view/?pid=34_3 |
| Summary: | The adoption of accents-based speech recognition into the e- learning environment has revolutionized the e-learning technology and reducing the learners’ barriers to knowledge acquisition, particularly as it relates to the effects of accents on the participants’ comprehension. Several researchers had worked on the development of accents-based speech recognition models using different techniques and architectures, such as acoustic model adaptation, pronunciation adaptation, Restricted Boltzmann Machine (RBM), Hidden Markow Model, Gaussian Mixture Model, Artificial Neural Network, Recurrent Neural Networks, Convolution Neural Network, among others. However, the accuracy rate of these models with these architectures and techniques have also become subject of research discussions. In this paper, we propose a new approach that combines the Long-Short Term Memory (LSTM) and Restricted Boltzmann Machine (RBM) of deep neural network techniques to form an optimal architecture for the accents-based speech recognition system development |
|---|---|
| ISSN: | 2066-4273 2066-3129 |