KRITIK INFRATUZILMALARDA MAXFIYLIKNI SAQLAYDIGAN SUN'IY INTELLEKT UCHUN KVANT-HIMOYALANGAN FEDERATIV TA'LIM ARXITEKTURASI
DOI:
https://doi.org/10.65164/e8f5y359Keywords:
Federativ ta'lim, post-kvant kriptografiyasi, mashinaviy ta'lim, kiberxavfsizlik, maxfiylikni saqlash, kvant hisoblashlariAbstract
Maʼlumotlarga asoslangan sunʼiy intellekt (AI) modellari va foydalanuvchi maxfiyligi oʻrtasidagi ziddiyat bugungi kunning asosiy muammolaridan biridir. Federativ taʼlim (FL) markazlashtirilmagan holda modellarni oʻqitish orqali bu muammoni hal qilishga qaratilgan istiqbolli yondashuv hisoblanadi. Biroq, FL tizimlarida model yangilanishlarini himoya qilish uchun qoʻllaniladigan klassik kriptografik protokollar yaqinlashib kelayotgan kvant hisoblashlari tahdidi oldida zaifdir. Ushbu tadqiqotda biz Q-Fed (Quantum-Resistant Federated Learning) deb nomlangan yangi arxitekturani taklif etamiz. Q-Fed federativ taʼlim jarayonini post-kvant kriptografiyasi (PQC) algoritmlari bilan integratsiya qilib, tizimni ham klassik, ham kvant hujumlaridan himoya qiladi. Biz panjaraga asoslangan CRYSTALS-Kyber (KEM) va CRYSTALSDilithium (imzo) sxemalarini qoʻllagan holda arxitekturani ishlab chiqdik va uning samaradorligini keng qamrovli simulyatsiyalar orqali tahlil qildik. Natijalar shuni koʻrsatadiki, Q-Fed model aniqligini pasaytirmagan holda kvant-barqaror xavfsizlikni taʼminlaydi, biroq hisoblash va kommunikatsiya xarajatlarida oʻlchanadigan, ammo boshqariladigan ortish kuzatiladi. Ushbu ish kelajak avlod AI tizimlarining maxfiyligi va xavfsizligini taʼminlash uchun amaliy va istiqbolli yechim taklif etadi
References
1. Shlezinger, A., & Shlezinger, M. (2020). The New AI: The Future of Artificial Intelligence is
Here. Apress.
2. McMahan, B., Moore, E., Ramage, D., Hampson, S., & y Arcas, B. A. (2017). CommunicationEfficient Learning of Deep Networks from Decentralized Data. Proceedings of the 20th
International Conference on Artificial Intelligence and Statistics (AISTATS).
3. Shor, P. W. (1994). Algorithms for quantum computation: discrete logarithms and factoring.
Proceedings of the 35th Annual Symposium on Foundations of Computer Science.
4. Alagic, G., et al. (2020). Status Report on the Second Round of the NIST Post-Quantum
Cryptography Standardization Process. NIST Internal Report 8309.
5. Li, T., Sahu, A. K., Zaheer, M., Sanjabi, M., Talwalkar, A., & Smith, V. (2020). Federated
Optimization in Heterogeneous Networks. Proceedings of the Conference on Machine Learning
and Systems (MLSys).
6. Bhagoji, A. N., Chakraborty, S., Mittal, P., & Calo, S. (2019). Analyzing Federated Learning
through Poisoning Attacks. arXiv preprint arXiv:1911.11835.
7. Zhu, L., Liu, Z., & Han, S. (2019). Deep Leakage from Gradients. Advances in Neural
Information Processing Systems (NeurIPS), 32.
8. Bernstein, D. J., & Lange, T. (2017). Post-quantum cryptography. Nature, 549(7671), 188-194.
9. Yin, X., et al. (2021). A Post-Quantum Secure Federated Learning Framework. 2021 IEEE
International Conference on Blockchain and Cryptocurrency (ICBC).
10. Ducas, L., Lyubashevsky, V., & Prest, T. (2018). CRYSTALS-Kyber: a CCA-secure modulelattice-based KEM. 2018 IEEE European Symposium on Security and Privacy (EuroS&P).
11. Ducas, L., Kiltz, E., Lepoint, T., Lyubashevsky, V., Schwabe, P., Schanck, G., & Stehle, D.
(2018). CRYSTALS-Dilithium: a lattice-based digital signature scheme. IACR Transactions on
Cryptographic Hardware and Embedded Systems, 2018(1), 238-268.
12. Kairouz, P., et al. (2021). Advances and Open Problems in Federated Learning. Foundations and
Trends in Machine Learning, 14(1–2), 1-210
