- Sirdaryo viloyati Guliston shahar
- Dushanba-Shanba: 9.00 - 18.00
Xaitbayev Azizbek Pirnazarovich
10.5281/zenodo.22938676
Annotatsiya
Generativ sun’iy intellekt xizmatlariga yuboriladigan prompt, hujjat va dialog kontekstlarida shaxsiy, moliyaviy, tibbiy, autentifikatsiya hamda tijoratga oid ma’lumotlar aralash holda keladi. Mazkur maqolada bunday ma’lumotlarni aniqlash usullari qoida-asosli filtrlar, regex va lug‘atlar, NER/NLP, klassik mashinaviy o‘qitish, chuqur o‘rganish, Transformer, LLM-asosli, gibrid, kontekstga va siyosatga asoslangan hamda lokal maxfiylikni saqlovchi tizimlar sifatida tasniflanadi. Tahlil shuni ko‘rsatadiki, regex va lug‘atlar yuqori tezlikda ishlasa-da, semantik kontekstni yomon ajratadi; Transformer va LLM usullari kontekstni yaxshiroq modellashtiradi, biroq kechikish, xarajat, deterministik bo‘lmagan chiqish va maxfiy promptni uchinchi tomonga yuborish xavfini keltiradi. Shuning uchun maqolada lokal kaskadli detektor taklif qilinadi: avval yuqori aniqlikdagi qoidalar, keyin kichik Transformer, noaniq holatlarda esa siyosat-kontekst skoringi qo‘llanadi. Taklif etilayotgan modelning formulalari va baholash protokoli empirik tekshiruv uchun beriladi; ular tayyor ilmiy natija emas.
Kalit so'zlar:
generativ sun’iy intellekt, LLM, prompt maxfiyligi, PII, NER, DLP,
kontekstga asoslangan aniqlash, false positive, false negative, lokal middleware.
Foydalanilgan adabiyotlar
1. Sun X, Liu G, He Z, et al (2024) DePrompt: Desensitization and Evaluation of Personal Identifiable Information in Large Language Model Prompts. arXivorg. https://doi.org/10.48550/arXiv.2408.08930 2. Jha P (2026) PIIBench: A Unified Multi-Source Benchmark Corpus for Personally Identifiable Information Detection. arXivorg. https://doi.org/10.5281/zenodo.19533263 3. Vats G, Agrawal A, Singhal S, et al (2026) REDACT: A Systematically Controlled Multilingual Benchmark for Personal Information Detection. arXivorg. https://doi.org/10.48550/arXiv.2606.19881 4. Savkin M, Ionov T, Konovalov V (2025) SPY: Enhancing Privacy with Synthetic PII Detection Dataset. North American Chapter of the Association for Computational Linguistics. https://doi.org/10.18653/v1/2025.naacl-srw.23 5. (2026) A Survey of Zero-Shot Sensitive Information Detection Techniques based on Large Language Models. Scientific Journal of Intelligent Systems Research. https://doi.org/10.54691/3kzsns60 6. Mainetti L, Elia A (2025) Detecting Personally Identifiable Information Through Natural Language Processing: A Step Forward. Applied System Innovation. https://doi.org/10.3390/asi8020055 7. Nokkaew P, Anekboon K (2025) Enhancing Personally Identifiable Information Recognition Using Compact Transformer With Hybrid Sequential Architecture. International journal of advances in signal and image sciences. https://doi.org/10.29284/m14da163 8. P N, Philip T, K. Y, Chandrika R (2026) Design and Implementation of an Automated PII Detection System using DeBERTaV2 and NLP. 2026 6th International Conference on Image Processing and Capsule Networks (ICIPCN). https://doi.org/10.1109/ICIPCN67432.2026.11438673