O‘ZBEK TILIDAGI TOVUSH VA MATINLARGA RAQAMLI ISHLOV BERISH ORQALI ULAR ORASIDAGI MOSLIK DARAJASINI ANIQLASH VA BAXOLASH ALGARITIMLARI.

Baxromov A. B.


Annotatsiya

So‘nggi yillarda sun’iy intellekt asosidagi nutqni matnga aylantirish (ASR) texnologiyalari tez sur’atlarda rivojlanib bormoqda. Bu jarayon, ayniqsa, ta’lim sohasida test savollariga og‘zaki javoblarni avtomatik tarzda yozib olish va baholash imkonini yaratadi. Tadqiqotning asosiy maqsadi — talaba yoki foydalanuvchining og‘zaki javobi ovoz yozuvi sifatida qabul qilinadi, keyin STT (Speech-to-Text) modellaridan foydalanib matnga aylantiriladi, hosil bo‘lgan matn esa etalon javob bilan solishtiriladi va avtomatik tarzda baholanadi. Baholash jarayonida klassik ko‘rsatkichlar — WER (Word Error Rate), CER (Character Error Rate) hamda zamonaviy semantik o‘xshashlik metrikalari qo‘llaniladi. Tadqiqot natijalari shuni ko‘rsatadiki, an’anaviy xatolik ko‘rsatkichlari bilan bir qatorda semantik o‘xshashlikni baholash foydalanuvchining javobini yanada adolatli tahlil qilish imkonini beradi. Ushbu yondashuv ta’limda test baholash tizimlarini avtomatlashtirish, shuningdek, apellyatsiya jarayonlarida talabalarning huquqlarini ta’minlashda samarali vosita bo‘lib xizmat qilishi mumkin.


Adabiyotlar

  1. Frederick Jelinek. “Statistical Methods for Speech Recognition”, Cambridge 1997.
  2. Lawrence Rabiner va Biing-Hwang Juang. nutq tanish texnologiyalarini. “A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition”, 1989.
  3. James L. Flanagan. Bell Labs. Speech Analysis, Synthesis and Perception 1972.
  4. Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, Yoshua Bengio. Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling 2014-yil.
  5. Zhiyuan Tang, Ying Shi, Dong Wang, Yang Feng, Shiyue Zhang. Memory Visualization for Gated Recurrent Neural Networks in Speech Recognition 2017.
  6. Andrew Ng va Baidu Research. Deep Speech: Scaling Up End-to-End Speech Recognition.” arXiv preprint arXiv 2014.
  7. Muhammadjon Musayev, Saida Musaxo‘jayeva, Ilyos Hujayorov, Yerbo‘lat Xasanov Mannon Ochilov va Huseyin Atakan Vorillar. An Open-Source Uzbek Speech Corpus and Initial Speech Recognition 2021.
  8. Abduqahhor Saidov, Sardorbek Yusupov, Anvar Narzullaev, Alisher Matyakubov, Saidmurod Radjabov, Hafiz Ibrahim, Jasur Temirov. A 60 Hour Uzbek Read Speech Corpus with Punctuation, Casing, and Context 2024.
  9. Mukhamadiyev, A.; Mukhiddinov, M.; Khujayarov, I.; Ochilov, M.; Cho,J. Development of Language Models for Continuous Uzbek Speech Recognition System 2023.
  10. Salayeva Maxliyo, Kuriyozov Elmurod, Salayev Ulug‘beklarning Uzbek automatic speech recognition models using deep learning techniques 2023.
  11. Thennal D K, Jesin James, Deepa P Gopinath, Muhammed Ashraf Klarning Advocating Character Error Rate for Multilingual ASR Evaluation
  12. Hila Gonen1 Shauli Ravfogen, Yanai Elazar, Yoav Goldberg, larning “It’s not Greek to mBERT: Inducing Word-Level Translations from Multilingual BERT November 2020.
  13. Mikhail Evtikhiev, Egor Bogomolov, Yaroslav Sokolov, Timofey Bryksin. Out of the BLEU: how should we assess quality of the Code Generation models? May 2023.
  14. Zhanibek Kozhirbayev. Kazakh Speech Recognition: Wav2vec2.0 vs. Whisper January 2023.
  15. Ibragimova S.N. Turayev B.Sh. Abdullayeva M.I.larning Solving the problems ofnormalization of non-standardwords in the text of the uzbeklanguage November r 2023.