| Author(s) | Collection number | Pages | Download abstract | Download full text |
|---|---|---|---|---|
| Liakh I. M., Кляп М. П., Ціпіньо А. Ю., Роман А. М. | № 2 (90) | 128-145 |
|
|
The article explores the evolutionary development of machine translation – from rule-based systems (RBMT) and statistical methods (SMT) to next-generation neural architectures (NMT). Particular attention is devoted to the comparative analysis of morphological, syntactic, and semantic distinctions between English and Ukrainian that determine the limitations of automated translation accuracy. Among the principal divergences identified are inflectional variability, differences in word order, grammatical gender, number, aspect, and agreement categories, all of which constitute linguistic barriers to the machine reproduction of natural Ukrainian syntax.
To address these challenges, a hybrid linguistically oriented translation model is proposed, integrating morphosyntactic analysis mechanisms with a neural network architecture. The article provides a detailed description of the system’s stepwise framework, which comprises a morphosyntactic normalization preprocessing block, a linguistic feature extraction module, a neural translation core, a Ukrainian morphogenerator, and a post-editing subsystem. For an objective evaluation of the proposed architecture’s performance, the implementation of international translation quality metrics – BLEU (Bilingual Evaluation Understudy) and TER (Translation Edit Rate) – is envisaged, enabling the assessment of translation accuracy and naturalness relative to human-generated reference outputs.
The authors substantiate the feasibility of incorporating grammatical and contextual rules as integral components of neural models designed for the translation of inflectional languages, particularly Ukrainian. This approach is aimed at enhancing the accuracy, stylistic naturalness, and contextual adequacy of computational translation systems. The research findings outline prospects for further development in constructing large-scale English–Ukrainian parallel corpora, experimentally validating the proposed model, and extending the approach to multilingual translation platforms. In practical terms, the study contributes to the advancement of high-quality Ukrainian-language translation tools integrated into contemporary natural language processing systems.
Keywords: automated translation; machine translation; neural network; morphosyntactic analysis; BLEU; TER.
doi: 10.32403/0554-4866-2025-2-90-128-145
- 1. Tomenchuk, M., & Popovych, K. (2024). Large Language Models and Machine Translation. Věda a Perspektivy. 11. 422–431. https://doi.org/10.52058/2695-1592-2024-11(42)-422-431.
- 2. Lytvyn, V., Pukach, P., Vysotska, V., Vovk, M., & Kholodna, N. (2023). Identification and correction of grammatical errors in Ukrainian texts based on machine learning technology. Mathematics, 11(4), 904. https://doi.org/10.3390/math11040904.
- 3. Mandziy, K. S., Yurlova, U. V., & Dilai, M. P. (2022). English-Ukrainian Parallel Corpus of IT Texts: Application in Translation Studies. In COLINS. 724-736. https://api.semanticscholar.org/CorpusID:250625296.
- 4. Vysotska, V. (2024). Computer linguistic system modelling for Ukrainian language processing. In CEUR Workshop Proceedings. 3722. 288-342. https://dblp.org/rec/conf/colins/Vysotska24b.bib.
- 5. Che, C. (2022). Application of Differential Evolution Algorithm in the Construction and Simulation of Interactive English Translation Teaching Mode. Advances in Multimedia, 2022(1), 6182551. https://doi.org/10.1155/2022/6182551.
- 6. Zhang, J., Luan, H., Sun, M., Zhai, F., Xu, J., & Liu, Y. (2021). Neural machine translation with explicit phrase alignment. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 29, 1001-1010. https://doi.org/10.1109/TASLP.2021.3057831.
- 7. Bhuvaneswari, K., & Varalakshmi, M. (2024). Efficient incremental training using a novel NMT-SMT hybrid framework for translation of low-resource languages. Frontiers in Artificial Intelligence, 7, 1381290. https://doi.org/10.3389/frai.2024.1381290.
- 8. Li, Z., Gao, S., Li, X., & Asghari, H. B. (2022, June). Strengthening Research of SYSTRAN in the Field of Artificial Intelligence Automatic Translation. In International Conference on Applications and Techniques in Cyber Intelligence (pp. 626-633). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-031-29097-8_74.
- 9. Soliman, A. S., Hadhoud, M. M., & Shaheen, S. I. (2022). MarianCG: a code generation transformer model inspired by machine translation. Journal of Engineering and Applied Science, 69(1), 104. https://doi.org/10.1186/s44147-022-00159-4.
- 10. Demenchuk, O. (2025). Semantic aspects of translation: English-Ukrainian correspondences. Naukovi zapysky Vinnytsʹkoho derzhavnoho pedahohichnoho universytetu imeni Mykhayla Kotsyubynsʹkoho. Seriya: Filolohiya (movoznavstvo), (40), (40), 141-147. https://doi.org/ 10.31652/2521-1307-2025-40-15.
- 11. Han, C. (2025). Quality assessment in multilingual, multimodal, and multiagent translation and interpreting (QAM3 T&I): Proposing a unifying framework for research. Interpreting and Society, 5(1), 27-55. https://doi.org/10.1177/27523810251322645.
- 12. Ponomarevskyi, A. (2022). Variability of artificial intelligence related terms used in translation from English into Ukrainian [Doctoral dissertation, Kauno technologijos universitetas]. Laisvai prieinamas internete / Unrestricted online access. https://epubl.ktu.edu/object/elaba: 132032271/.
- 13. Castilho, S., & Knowles, R. (2025). A survey of context in neural machine translation and its evaluation. Natural Language Processing, 31(4), 986-1016. https://doi.org/10.1017/nlp.2024.7.
