Elamur Research⌕
☰
home / papers / 2025 / Universal Dependencies Treebank for Uzbek

Universal Dependencies Treebank for Uzbek

Summary

Uzbek belongs to the Turkic family and has more than 40 million speakers, yet few annotated resources exist for it. In particular, there has been no gold-standard syntactically annotated corpus. The authors created the first treebank for Uzbek within Universal Dependencies, named Uzbek-UT. It comprises 500 sentences in Latin script: half from news texts in the UzCrawl dataset and half from 20th–21st-century fiction. Annotation was semi-automatic. Tokenization was performed with NLTK, and lemmatization and initial POS tagging with the UzMorphAnalyzer tool. Morphological features and dependencies were first annotated manually on a small sample. Stanza models were trained on this sample, and a native speaker with linguistic training corrected their predictions. The cycle was repeated three times, and the automatic models became more accurate with each round. The authors also examine Uzbek constructions that fit poorly into UD: verbal multiword expressions, multiple voice affixes on a single verb, and postverbial constructions. Despite its small size, the corpus provides a basis for training taggers and parsers and for comparative research on Turkic languages.

Key points

  • Uzbek-UT contains 500 sentences (5850 tokens, 3523 unique word forms). The corpus is annotated with 17 UPOS tags, 42 morphological features and 32 dependency types. The XPOS and DEPS fields were left empty.
  • The corpus is openly available in the UD_Uzbek-UT repository on GitHub.
  • The Stanza parser with fastText vectors was trained three times on progressively larger samples of 100, 240 and 400 sentences. F1 rose from 46.26 to 52.43 for dependencies, from 69.39 to 73.08 for UPOS, and from 86.78 to 92.11 for lemmatization.
  • Postverbial constructions (e.g., to‘xtab qol) are annotated inconsistently in other Turkic UD treebanks: via xcomp in the Uyghur treebank and via aux in the Kyrgyz one. The authors annotate them with the compound relation, with the postverbial verb as the head, and propose introducing a dedicated subtype of this relation.
  • In expressions such as tashkil qil, the first element does not exist as a standalone word, so it is tagged as PART. When a verb has several voice affixes, the Voice value is determined by the outermost one.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Universal Dependencies · Treebanks

Cite

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{akhundjanova2025universal,
  title = {{Universal Dependencies Treebank for Uzbek}},
  author = {Akhundjanova, Arofat and Talamo, Luigi},
  booktitle = {Proceedings of the Third Workshop on Resources and Representations for Under-Resourced Languages and Domains (RESOURCEFUL-2025)},
  year = {2025},
  url = {https://aclanthology.org/2025.resourceful-1.1/}
}

References

24 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Alessandro Agostini, Timur Usmanov, Ulugbek Kham-damov, Nilufar Abdurakhmonova, and Mukhammadsaid Mamasaidov. 2021. UZWORDNET: A lexical-semantic database for the Uzbek language. In Proceedings of the 11th Global Wordnet Conference, pages 8–19, University of South Africa (UNISA). Global Wordnet Asso… find
  2. Edward Loper and Steven Bird. 2002. NLTK: the natural language toolkit. In Proceedings of the ACL-02 Workshop on Effective tools and methodologies for teaching natural language processing and computational linguistics, volume 1, page 63–70. Association for Computational Linguistics. find
  3. Hendrik Boeschoten. 2021a. The speakers of Turkic languages. In Lars Johanson and Éva Á. Csató, editors, The Turkic languages, pages 1–12. Routledge. find
  4. Kh. A. Madatov, D. J. Khujamov, and B. R. Boltayev. 2022. Creating of the Uzbek WordNet based on Turkish WordNet. In AIP Conference Proceedings, volume 2432. AIP Publishing. find
  5. Hendrik Boeschoten. 2021b. Uzbek. In Lars Johanson and Éva Á. Csató, editors, The Turkic languages, pages 388–408. Routledge. find
  6. Mukhammadsaid Mamasaidov and Abror Shopulatov. 2023. Uzcrawl dataset. find
  7. B. Mansurov and A. Mansurov. 2021. UzBERT: pretraining a BERT model for Uzbek. CoRR, abs/2108.09814. — [arXiv:2108.09814](https://arxiv.org/abs/2108.09814) find
  8. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,… find
  9. Marie-Catherine de Marneffe, Christopher D. Manning, Joakim Nivre, and Daniel Zeman. 2021. Universal Dependencies. Computational Linguistics, 47(2):255–308. find our card
  10. Gayrat Matlatipov and Zygmunt Vetulani. 2009. Representation of Uzbek Morphology in Prolog, page 83–110. Springer-Verlag, Berlin, Heidelberg. find
  11. Edouard Grave, Piotr Bojanowski, Prakhar Gupta, Armand Joulin, and Tomas Mikolov. 2018. Learning word vectors for 157 languages. In Proceedings of the International Conference on Language Resources and Evaluation (LREC 2018). find our card
  12. Sanatbek Matlatipov, Hulkar Rahimboeva, Jaloliddin Rajabov, and Elmurod Kuriyozov. 2022. Uzbek sentiment analysis based on local restaurant reviews. In Proceedings of the ALTNLP The International Conference and workshop on Agglutanative Language Technologies as a challenge of Natural Language Proces… find
  13. Bruno Guillaume. 2021. Graph matching and graph rewriting: GREW tools for corpus exploration, maintenance and conversion. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations, pages 168–175, Online. Association for Comp… find
  14. Peng Qi, Yuhao Zhang, Yuhui Zhang, Jason Bolton, and Christopher D. Manning. 2020. Stanza: A python natural language processing toolkit for many human languages. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations. Association for Computa… find our card
  15. Lars Johanson. 2021. The structure of Turkic. In Lars Johanson and Éva Á. Csató, editors, The Turkic languages, pages 26–59. Routledge. find
  16. Elmurod Kuriyozov, Sanatbek Matlatipov, Miguel A. Alonso, and Carlos Gómez-Rodrı́guez. 2019. Construction and evaluation of sentiment datasets for low-resource languages: The case of Uzbek. In Human Language Technology. Challenges for Computer Science and Linguistics: 9th Language and Technology Co… find
  17. I. M. Rabbimov and S. S. Kobilov. 2020. Multi-class text classification of Uzbek news articles using machine learning. Journal of Physics: Conference Series, 1546(1):012097. find
  18. Shavkat Rahmatullayev. 2006. Hozirgi Adabiy O‘zbek Tili [Contemporary Literary Uzbek Language (textbook)]. Universitet. find
  19. Ulugbek Salaev. 2023. Modeling morphological analysis based on word-ending for Uzbek language. Science and innovation, 2(C11):29–34. find
  20. Ulugbek Salaev, Elmurod Kuriyozov, and Carlos Gómez-Rodrı́guez. 2022. SimRelUz: Similarity and relatedness scores as a semantic evaluation dataset for Uzbek language. In Proceedings of the 1st Annual Meeting of the ELRA/ISCA Special Interest Group on Under-Resourced Languages, pages 199– 206, Marse… find
  21. Maksud Sharipov, Elmurod Kuriyozov, Ollabergan Yuldashev, and Ogabek Sobirov. 2023. Uzbektagger: The rule-based POS tagger for Uzbek language. arXiv preprint arXiv:2301.12711. find
  22. Maksud Sharipov, Jamolbek Mattiev, Jasur Sobirov, and Rustam Baltayev. 2022. Creating a morphological and syntactic tagged corpus for the Uzbek language. In Proceedings of the ALTNLP The International Conference and workshop on Agglutanative Language Technologies as a challenge of Natural Language P… find
  23. Maksud Sharipov and Ogabek Sobirov. 2022. Development of a rule-based lemmatization algorithm through finite state machine for Uzbek language. In Proceedings of the ALTNLP The International Conference and workshop on Agglutanative Language Technologies as a challenge of Natural Language Processing,… find
  24. Maksud Sharipov and Ollabergan Yuldashov. 2022. Uzbekstemmer: Development of a rule-based stemming algorithm for Uzbek language. In Proceedings of the ALTNLP The International Conference and workshop on Agglutanative Language Technologies as a challenge of Natural Language Processing, Virtual Event,… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/akhundjanova-2025-universal-dependencies-treebank-uzbek/