AUTOMATED MULTI-DIALECT SPEECH RECOGNITION USING STACKED ATTENTION-BASED DEEP LEARNING WITH NATURAL LANGUAGE PROCESSING MODEL

Alanoud Al Mazroa, Achraf Ben Miled, Mashael M Asiri, Yazeed Alzahrani, Ahmed Sayed and Faisal Mohammed Nafie
Additional contact information
Alanoud Al Mazroa: Department of Information Systems, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman University, P. O. Box 84428, Riyadh 11671, Saudi Arabia
Achraf Ben Miled: ��Department of Computer Science, College of Science, Northern Border University, Arar 73213, Saudi Arabia
Mashael M Asiri: ��Department of Computer Science, Applied College at Mahayil, King Khalid University, Abha 62521, Saudi Arabia
Yazeed Alzahrani: �Department of Computer Engineering, College of Engineering in Wadi Addawasir, Prince Sattam Bin Abdulaziz University, Al-Kharj 16278, Saudi Arabia
Ahmed Sayed: �Research Center, Future University in Egypt, New Cairo 11835, Egypt
Faisal Mohammed Nafie: ��Department of Natural and Applied Sciences, Community College, Majmaah University, Al Majma’ah 11952, Saudi Arabia

FRACTALS (fractals), 2024, vol. 32, issue 09n10, 1-13

Abstract: Dialects are language variations that occur due to differences in social groups or geographical regions. Dialect speech recognition is the approach to accurately transcribe spoken language that involves regional variation in vocabulary, syntax, and pronunciation. Models need to be trained on various dialects to handle linguistic differences effectively. The latest advancements in automatic speech recognition (ASR) and complex systems methods are showing progress in recurrent neural networks (RNN), deep neural networks (DNN), and convolutional neural networks (CNN). Multi-dialect speech recognition remains a challenge, notwithstanding the progress of deep learning (DL) in speech recognition for many computing applications in environmental modeling and smart cities. Even though the dialect-specific acoustic model is known to perform well, it is not easier to maintain when the number of dialects for all the languages is large and dialect-specific data are limited. This paper offers an Automated Multi-Dialect Speech Recognition using the Stacked Attention-based Deep Learning (MDSR-SADL) technique in environmental modeling and smart cities. The MDSR-SADL technique primarily applies the DL model to identify various dialects. In the MDSR-SADL technique, stacked long short-term memory with attention-based autoencoder (SLSTM-AAE) model is used, which integrates stack modeling with LSTM and AE. Besides, the attention model enables dialect identification by offering dialect details for speech identification. The MDSR-SADL model uses the Fractals Harris Hawks Optimization (FHHO) model for hyperparameter selection. A sequence of simulations was implemented to illustrate the improved solution of the MDSR-SADL model. The experimental investigation of the MDSR-SADL technique exhibits superior accuracy values of 99.52% and 99.55% over other techniques under Tibetan and Chinese datasets.

Keywords: Automatic Speech Recognition; Artificial Intelligence; Complex Systems; Fractals Harris Hawks Optimization; Deep Learning; Dialect Identification (search for similar items in EconPapers)
Date: 2024
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.worldscientific.com/doi/abs/10.1142/S0218348X25400304
Access to full text is restricted to subscribers

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:wsi:fracta:v:32:y:2024:i:09n10:n:s0218348x25400304

Ordering information: This journal article can be ordered from

DOI: 10.1142/S0218348X25400304

Access Statistics for this article

FRACTALS (fractals) is currently edited by Tara Taylor

More articles in FRACTALS (fractals) from World Scientific Publishing Co. Pte. Ltd.
Bibliographic data for series maintained by Tai Tone Lim ().