ANALYSIS OF PARAPHRASE GENERATION ALGORITHMS (ON THE EXAMPLE OF UZBEK, ENGLISH, AND RUSSIAN LANGUAGES)
This article presents a comparative analysis of paraphrase generation algorithms using Uzbek, English, and Russian languages as case studies. The paper focuses on seq2seq, Transformer-based (BERT, T5), and multilingual models (mBART, XLM-R). It evaluates how each language's morphological and syntactic structure impacts paraphrasing. The study highlights challenges specific to the agglutinative structure of the Uzbek language and offers transfer learning as a solution.
Chen, Q., Zhu, X., Ling, Z., Wei, S., Jiang, H. (2018). Enhancing Sentence Embedding with Generalized Pooling. ACL, pp. 448–457.
2. Mallinson, J., Sennrich, R., Lapata, M. (2017). Paraphrasing Revisited with Neural Machine Translation. EACL, pp. 400–410.
3. Vaswani, A., Shazeer, N., Parmar, N., et al. (2017). Attention Is All You Need. NeurIPS, pp. 5999–6009.
4. Raffel, C., Shazeer, N., Roberts, A., et al. (2020). Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer. JMLR, 21(140), pp. 1–67.
5. Al-Shameri, F., Al-Khalifa, H. (2024). Paraphrasing Approaches for Arabic and Related Morphologically Rich Languages. LREC, p. 6.
6. Prakash, A., Hasan, S. A., Lee, K., et al. (2016). Neural Paraphrase Generation with Stacked Residual LSTM Networks. COLING, pp. 1243–1252.
7. Sharipov, S. (1978). O‘zbek tilining grammatik qurilishi. Toshkent: Fan.
8. Rahmatullayev, A. (2022). O‘zbek tilining morfologik xususiyatlari va kompyuter lingvistikasi. Tashkent: O‘zMU nashriyoti.
9. Gerasimov, V. (2015). Zadachi Voyennoy Nauki, Obzor voyenno-politicheskoy obstanovki v mire. http://dokwar.ru/publ/novosti_i_sobytija/mnenie/zadachi_voennoj_nauki/9-1-0-681 (retrieved 02.02.2015)
10. Carr, J. (2012). Inside Cyber Warfare. Mapping the Cyber Underworld. O’Reilly, 2nd edition.
11. Çelikyilmaz, A., Bosselut, A., He, X., Choi, Y. (2021). Deep Language Generation for Agglutinative Languages: The Case of Turkish. ACL Anthology, pp. 88–96.
Copyright (c) 2025 «ACTA NUUz»
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
.jpg)
1.png)
