Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

GraphUnet-SS: A novel deep learning model for protein secondary structure prediction based on U-Net architecture.

Computers in biology and medicine·2026
Same author

HearteXplain: explainable prediction of acute heart failure and identification of hematologic biomarkers using EBMs and Morris sensitivity analysis.

Scientific reports·2025
Same author

Machine Learning Classification of Cognitive Status in Community-Dwelling Sarcopenic Women: A SHAP-Based Analysis of Physical Activity and Anthropometric Factors.

Medicina (Kaunas, Lithuania)·2025
Same author

Prediction of obesity levels based on physical activity and eating habits with a machine learning model integrated with explainable artificial intelligence.

Frontiers in physiology·2025
Same author

Precision Enhanced Bioactivity Prediction of Tyrosine Kinase Inhibitors by Integrating Deep Learning and Molecular Fingerprints Towards Cost-Effective and Targeted Cancer Therapy.

Pharmaceuticals (Basel, Switzerland)·2025
Same author

Biomarker discovery and development of prognostic prediction model using metabolomic panel in breast cancer patients: a hybrid methodology integrating machine learning and explainable artificial intelligence.

Frontiers in molecular biosciences·2025

Related Experiment Video

Updated: Jun 27, 2025

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
05:48

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception

Published on: August 9, 2024

1.5K

Customized deep learning based Turkish automatic speech recognition system supported by language model.

Yasin Görmez1

  • 1Management Information System, Sivas Cumhuriyet University, Sivas, Merkez, Turkiye.

Peerj. Computer Science
|April 25, 2024
PubMed
Summary
This summary is machine-generated.

This study introduces deep learning models for Turkish automatic speech recognition, significantly improving accuracy by integrating a language model. The developed system shows superior performance compared to existing literature.

Keywords:
Automatic speech recognitionDeep learningMachine learningSequence to sequence modelTurkish speech recognationWord normalization

More Related Videos

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody
09:09

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Published on: September 27, 2024

443
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

548

Related Experiment Videos

Last Updated: Jun 27, 2025

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception
05:48

Author Spotlight: Investigating the Impact of Emotional Prosodies on Voice Recognition and Perception

Published on: August 9, 2024

1.5K
Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody
09:09

Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody

Published on: September 27, 2024

443
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
03:14

Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness

Published on: December 6, 2024

548

Area of Science:

  • Natural Language Processing
  • Machine Learning
  • Speech Technology

Background:

  • Automatic speech recognition (ASR) systems are crucial for daily applications but lag in less common languages like Turkish.
  • Turkish's agglutinative structure poses unique challenges for ASR system development.
  • Existing ASR research has not sufficiently addressed the needs of the Turkish language.

Purpose of the Study:

  • To propose advanced deep learning models for Turkish automatic speech recognition.
  • To enhance ASR performance for Turkish through language model integration.
  • To address the specific linguistic challenges of Turkish in ASR.

Main Methods:

  • Developed deep learning models incorporating convolutional neural networks, GRUs, LSTMs, and transformer layers.
  • Utilized the Zemberek library to construct a language model for performance enhancement.
  • Applied Bayesian optimization for fine-tuning deep learning model hyper-parameters.
  • Evaluated performance using standard metrics: word error rate (WER) and character error rate (CER).

Main Results:

  • Without a language model, WER/CER scores were 22.2/14.05 (Turkish Microphone Speech Corpus) and 11.5/4.15 (Turkish Speech Corpus).
  • With language model integration, scores improved to 9.85/5.35 (Turkish Microphone Speech Corpus) and 8.4/2.7 (Turkish Speech Corpus).
  • The proposed models demonstrated superior performance over existing literature benchmarks.

Conclusions:

  • Deep learning models, enhanced by language modeling and optimized hyper-parameters, significantly improve Turkish ASR accuracy.
  • The developed approach offers a robust solution for ASR in morphologically rich languages like Turkish.
  • This research contributes to advancing speech recognition technology for under-resourced languages.