REVIEWER PROCESSES AND PARALLEL DICTIONARY CREATION IN ANNOTATING UZBEK DIALECTS FOR TRAINING ARTIFICIAL INTELLIGENCE MODELS ON THE EXAMPLE OF THE KHOREZM DIALECT

Authors

  • Shakhodat Qodirberganova Department of Modern information technologies Uzbekistan state world languages university Author

Keywords:

Artificial Intelligence (AI), Uzbek language, dialect annotation, reviewer process, Khorezm dialect, Natural Language Processing (NLP).

Abstract

This study presents the methodology for annotating the Khorezm dialect, a low-resource variety of the Uzbek language, and developing parallel dictionaries for training artificial intelligence (AI) models. The Khorezm dialect differs significantly from Standard Uzbek in its phonetic, lexical, and morphological features and, as a representative of the Oghuz dialect group, exhibits considerable linguistic affinity with the Turkish language. The article examines the role of reviewer processes in improving dataset quality and analyzes the impact of parallel dictionaries on the performance of Neural Machine Translation (NMT) systems. The findings highlight the importance of systematic review procedures and high-quality parallel lexical resources for enhancing the effectiveness of AI-based language technologies for dialectal Uzbek.

References

1.Ölmez, Mehmet. (2018). Türk Dilleri ve Lehçeleri. Ankara: Akçağ Yayınları.

2.Bender, Emily M.. (2019). The data statements for natural language processing: Toward mitigating system bias and enabling better science. Transactions of the Association for Computational Linguistics, 7, 587–604.

3.Zampieri, Marcos, et al. (2020). Language identification and dialect classification in low-resource settings. Computer Speech & Language, 62, 101088.

4.Yusupov, Erkin. (2019). Xorazm shevalarining fonetik va leksik xususiyatlari. O‘zbek Tili va Adabiyoti, 5, 45–53

Downloads

Published

2026-06-03