JSMA VA C&W YONDASHUVLARIGA ASOSLANGAN ZARARKUNANDA DASTURLARNI ANIQLASHDA ADVERSARIAL HUJUMINI YARATISH ALGORITMI
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Ushbu maqolada mashinali o'qitishga asoslangan zararkunanda dasturlarni aniqlash tizimlariga qarshi yaratiladigan adversarial hujumlarning ikkita usuli — JSMA (Jacobian Saliency Map Attack) va Carlini & Wagner (C&W) — amaliy kontekstda tahlil qilingan va ularni konvolyutsion (CNN) hamda ketma-ketlikka asoslangan (RNN/LSTM) klassifikatorlarga moslashtirish usullarini taklif etilgan. Tadqiqotning maqsadi — mavjud detektorlarning zaif tomonlarini identifikatsiya qilish, hujum senariylarini generatsiya qilish algoritmlarini takomillashtirish va shu asosda mustahkamroq himoya mexanizmlarini ishlab chiqish. Metodologiya qismi adversarial namunalar yaratish jarayonini matematik jihatdan ifodalaydi, JSMA va C&W algoritmlarining o'ziga xos modifikatsiyalari (xususan, diskret xususiyatli dastur kodlari uchun perturbatsiya cheklovlari va semantik yaxlitlikni saqlash mexanizmlari) batafsil tavsiflangan.
Fields of Study & Subjects
Adversarial Robustness in Machine Learning
Security and Verification in Computing
Software Testing and Debugging Techniques
Publication details
- Journal / source
- CyberLeninK (CyberLeninka)
- Publication date
- 27 March 2026
- Language
- UZ