Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic]
This paper presents the development of a speech recognition system for the Arabic language that can handle continuous speech and a large number of words, independent of the speaker, using deep neural network models trained by self-supervised learning. The system was built using the HuBERT model, and...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | Arabic |
| Published: |
Higher Commission for Scientific Research
2025-01-01
|
| Series: | Syrian Journal for Science and Innovation |
| Subjects: | |
| Online Access: | https://journal.hcsr.gov.sy/archives/1523 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832586218474307584 |
|---|---|
| author | Rima Sbih Assef Jafar Ali Kazem |
| author_facet | Rima Sbih Assef Jafar Ali Kazem |
| author_sort | Rima Sbih |
| collection | DOAJ |
| description | This paper presents the development of a speech recognition system for the Arabic language that can handle continuous speech and a large number of words, independent of the speaker, using deep neural network models trained by self-supervised learning. The system was built using the HuBERT model, and resulted in a word error rate (WER) of 19.3%. Our study on different data sets revealed that the HuBERT-based system has a significant ability to generalize to different spoken dialects. Additionally, we conducted a statistical analysis on the errors specific to the Arabic language that arise from the HuBERT-based system, which highlighted the necessity of incorporating an error correction language model to enhance system accuracy. After the addition of an Arabic language model, the WER decreased to 10.7%. Overall, this study emphasizes the potential of self-supervised learning-based speech recognition systems for the Arabic language and highlights the importance of incorporating language models to enhance system accuracy. |
| format | Article |
| id | doaj-art-ae432e68dd2b47f59c0139ba6aa1bec1 |
| institution | Kabale University |
| issn | 2959-8591 |
| language | Arabic |
| publishDate | 2025-01-01 |
| publisher | Higher Commission for Scientific Research |
| record_format | Article |
| series | Syrian Journal for Science and Innovation |
| spelling | doaj-art-ae432e68dd2b47f59c0139ba6aa1bec12025-01-26T08:24:57ZaraHigher Commission for Scientific ResearchSyrian Journal for Science and Innovation2959-85912025-01-013110.5281/zenodo.14723614Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic]Rima Sbih0Assef Jafar1Ali Kazem2Higher Institute for Applied Sciences and Technology_Damascus_Syria.Higher Institute for Applied Sciences and Technology_Damascus_Syria.Higher Institute for Applied Sciences and Technology_Damascus_Syria.This paper presents the development of a speech recognition system for the Arabic language that can handle continuous speech and a large number of words, independent of the speaker, using deep neural network models trained by self-supervised learning. The system was built using the HuBERT model, and resulted in a word error rate (WER) of 19.3%. Our study on different data sets revealed that the HuBERT-based system has a significant ability to generalize to different spoken dialects. Additionally, we conducted a statistical analysis on the errors specific to the Arabic language that arise from the HuBERT-based system, which highlighted the necessity of incorporating an error correction language model to enhance system accuracy. After the addition of an Arabic language model, the WER decreased to 10.7%. Overall, this study emphasizes the potential of self-supervised learning-based speech recognition systems for the Arabic language and highlights the importance of incorporating language models to enhance system accuracy.https://journal.hcsr.gov.sy/archives/1523speech recognitiondeep learningself-attentionsupervised learningself-supervised learning. |
| spellingShingle | Rima Sbih Assef Jafar Ali Kazem Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic] Syrian Journal for Science and Innovation speech recognition deep learning self-attention supervised learning self-supervised learning. |
| title | Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic] |
| title_full | Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic] |
| title_fullStr | Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic] |
| title_full_unstemmed | Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic] |
| title_short | Building Arabic Speech Recognition System Using HuBERT Model and Studying the Sources of Errors [Arabic] |
| title_sort | building arabic speech recognition system using hubert model and studying the sources of errors arabic |
| topic | speech recognition deep learning self-attention supervised learning self-supervised learning. |
| url | https://journal.hcsr.gov.sy/archives/1523 |
| work_keys_str_mv | AT rimasbih buildingarabicspeechrecognitionsystemusinghubertmodelandstudyingthesourcesoferrorsarabic AT assefjafar buildingarabicspeechrecognitionsystemusinghubertmodelandstudyingthesourcesoferrorsarabic AT alikazem buildingarabicspeechrecognitionsystemusinghubertmodelandstudyingthesourcesoferrorsarabic |