JBT Journal
Advanced

Raqamli Transformatsiya va Sun’iy Intellekt

O‘ZBEK TILIDAGI KONFIDENSIAL MATNLARNI ANIQLASH UCHUN TRANSFORMER MODELLAR VA GIBRID TASNIFLASH MEXANIZMLARIGA ASOSLANGAN DLP TIZIMI

Bozorov Obidjon Norqobilovich1

1 Mirzo Ulug‘bek nomidagi O‘zbekiston Milliy universiteti

Zenodo DOI: 10.5281/zenodo.20916240

Keywords

NER mashinali o‘qitish DLP tizimi matn tasniflash DistilBERT ML.NET konfidensial axborot ONNX

Abstract

Ushbu maqolada konfidensial axborotlarning ruxsatsiz chiqib ketishini (data leakage) aniqlash uchun mashinali o‘qitishga asoslangan matn tasniflash moduli va uning asosida yaratilgan “KARCHKOOT” DLP (Data Loss Prevention) tizimi taqdim etilgan. Maqolaning ilmiy yangiligi o‘zbek tili uchun qoida, NER va transformer modelini birlashtiruvchi gibrid tasniflash arxitekturasi hamda soxta salbiy holatlarni kamaytirishga qaratilgan ko‘p bosqichli qaror qabul qilish mexanizmini ishlab chiqishdan iborat. Tasniflash moduli ko‘p tilli DistilBERT modelini o‘zbek tilidagi konfidensial hujjatlar korpusida fine-tuning qilish va ONNX orqali C#/.NET muhitiga integratsiya qilish yo‘li bilan qurilgan. Miqdoriy baholash binar (konfidensial / ochiq) tasniflash masalasi sifatida 1 321 ta test faylida o‘tkazilgan; tizim 94,7% Accuracy, 91,5% Precision, 94,8% Recall va 93,1% F1-mezon ko‘rsatkichlariga erishdi. Ablation tahlili, baseline modellar bilan qiyoslash, obfuskatsiyaga chidamlilik va tahdid modeli tahlili keltirilgan. Tizim bir qator davlat tashkilotlarida amaliyotga tatbiq etilgan.

Views: 121 Published: 2026-04-28 Pages: 149-158 Volume: 4 Issue: 2

References

  1. Sanh V., Debut L., Chaumond J., Wolf T. DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter // arXiv:1910.01108. 2019.
  2. Yadav V., Bethard S. A Survey on Recent Advances in Named Entity Recognition from Deep Learning Models // Proceedings of COLING. 2018. P. 2145-2158.
  3. Liu Y. et al. RoBERTa: A Robustly Optimized BERT Pretraining Approach // arXiv:1907.11692. 2019.
  4. Ahmed Z. et al. Machine Learning at Microsoft with ML.NET // Proceedings of the 25th ACM SIGKDD. 2019. P. 2448-2458.
  5. ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator [Elektron resurs]. URL: https://onnxruntime.ai (murojaat sanasi: 10.11.2025).
  6. Hart M., Manadhata P., Johnson R. Text Classification for Data Loss Prevention // Privacy Enhancing Technologies Symposium (PETS). 2011. P. 18-37.
  7. Mansurov B., Mansurov A. UzBERT: pretrained model for the Uzbek language // arXiv:2108.09814. 2021.
  8. Gomez-Hidalgo J. M., Martin-Abreu J. M., Nieves J., Santos I., Brezo F., Bringas P. G. Data Leak Prevention through Named Entity Recognition // IEEE Int. Conf. on Social Computing (SocialCom/PASSAT). 2010. P. 1129-1134.
  9. Salaev U., Kuriyozov E., Gomez-Rodriguez C. SimRelUz: Similarity and Relatedness Scores as a Semantic Evaluation Dataset for Uzbek Language // Proceedings of the 1st Annual Meeting of the ELRA/ISCA SIG on Under-resourced Languages (SIGUL). 2022. P. 56-63.
  10. Shabtai A., Elovici Y., Rokach L. A Survey of Data Leakage Detection and Prevention Solutions. Springer, 2012. 92 p.
  11. Joachims T. Text Categorization with Support Vector Machines: Learning with Many Relevant Features // Proceedings of ECML. 1998. P. 137-142.
  12. Kim Y. Convolutional Neural Networks for Sentence Classification // Proceedings of EMNLP. 2014. P. 1746-1751.
  13. Conneau A. et al. Unsupervised Cross-lingual Representation Learning at Scale (XLM-R) // Proceedings of ACL. 2020. P. 8440-8451.
  14. Cohen J. A Coefficient of Agreement for Nominal Scales // Educational and Psychological Measurement. 1960. Vol. 20, No 1. P. 37-46.
  15. Salton G., Buckley C. Term-weighting approaches in automatic text retrieval // Information Processing & Management. 1988. Vol. 24, No 5. P. 513-523.
  16. Lample G., Ballesteros M., Subramanian S., Kawakami K., Dyer C. Neural Architectures for Named Entity Recognition // Proceedings of NAACL-HLT. 2016. P. 260-270.
  17. Dietterich T. G. Approximate Statistical Tests for Comparing Supervised Classification Learning Algorithms // Neural Computation. 1998. Vol. 10, No 7. P. 1895-1923.
  18. Vaswani A. et al. Attention Is All You Need // Advances in Neural Information Processing Systems (NeurIPS). 2017. Vol. 30.
  19. Devlin J., Chang M.-W., Lee K., Toutanova K. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding // Proceedings of NAACL-HLT. 2019. P. 4171-4186.
  20. Pires T., Schlinger E., Garrette D. How Multilingual is Multilingual BERT? // Proceedings of ACL. 2019. P. 4996-5001.
  21. Alneyadi S., Sithirasenan E., Muthukkumarasamy V. A survey on data leakage prevention systems // Journal of Network and Computer Applications. 2016. Vol. 62. P. 137-152.
  22. Liu S., Kuhn R. Data Loss Prevention // IT Professional. 2010. Vol. 12, No 2. P. 10-13.