Виды атак на федеративные нейросети и способы защиты

Федеративное обучение — это технология обучения с сохранением конфиденциальности в распределенных системах хранения данных. Такое обучение позволяет создать общую модель прогнозирования, сохраняя все данные в своих системах хранения. В обучении общей модели учувствуют несколько устройств, при этом каждое устройство имеет свои уникальные данные, на которых обучается нейросеть. Взаимодействие устройств происходит только для корректировки весов общей модели. После чего, обновленная модель передается на все устройства. Обучение на нескольких устройствах рождает множество возможностей для атак на этот тип сетей.

1. Kairouz, Peter; Brendan McMahan, H.; Avent, Brendan; Bellet, Aurélien; Bennis, Mehdi; Arjun Nitin Bhagoji; Bonawitz, Keith; Charles, Zachary; Cormode, Graham; Cummings, Rachel; D'Oliveira, Rafael G. L.; Salim El Rouayheb; Evans, David; Gardner, Josh; Garrett, Zachary; Gascón, Adrià; Ghazi, Badih; Gibbons, Phillip B.; Gruteser, Marco; Harchaoui, Zaid; He, Chaoyang; He, Lie; Huo, Zhouyuan; Hutchinson, Ben; Hsu, Justin; Jaggi, Martin; Javidi, Tara; Joshi, Gauri; Khodak, Mikhail; et al. (10 December 2019). "Advances and Open Problems in Federated Learning". arXiv:1912.04977, DOI: 10.48550/arXiv.1912.04977.

2. Federated Learning: Collaborative Machine Learning without Centralized Training Data (online) https://ai.googleblog.com/2017/04/federated-learning-collaborative.html — 01.12.2023.

3. Threats, attacks and defenses to federated learning: issues, taxonomy and perspectives - Peng Liu, Xiangru Xu, Wen Wang, Cybersecurity 5, 4 (2022), DOI: 10.1186/s42400-021-00105-6.

4. Aniruddha Saha; Akshayvarun Subramanya; Hamed Pirsiavash; (2019), - “Hidden Trigger Backdoor Attacks” - arXiv:1910.00033v2, DOI: 10.48550/arXiv.1910.00033.

5. Bhagoji AN, Chakraborty S, Mittal P, Calo SB - Analyzing federated learning through an adversarial lens. In: Chaudhuri K, Salakhutdinov R (eds) Proceedings of the 36th international conference on machine learning, ICML 2019, 9–15 June 2019, Long Beach, California, USA, volume 97 of proceedings of machine learning research, pp 634–643, PMLR 97:634-643, 2019.

6. Chen, X., Liu, C., Li, B., Lu, K., and Song, D. Targeted backdoor attacks on deep learning systems using data poisoning, arXiv:1712.05526, 2017a, DOI: 10.48550/arXiv.1712.05526.

7. Clement Fung, Chris J. M. Yoon, and Ivan Beschastnikh. The limitations of federated learning in sybil settings. In 23rd International Symposium on Research in Attacks, Intrusions and Defenses (RAID 2020), pp. 301–316, San Sebastian, October 2020. USENIX Association. ISBN 978-1-939133-18-2.

8. Naseri M, Hayes J, De Cristofaro E (2020) Toward robustness and privacy in federated learning: experimenting with local and central differential privacy. CoRR arXiv:2009.03561, DOI: 10.48550/arXiv.2009.03561

9. Abadi M, Chu A, Goodfellow I, McMahan HB, Mironov I, Talwar K, Zhang L (2016) Deep learning with differential privacy. In: Proceedings of the 2016 ACM SIGSAC conference on computer and communications security, pp 308–318, DOI: 10.48550/arXiv.1607.00133.

10. Kostenko V.A., Tankaev I.R., Federated Learning Using Simple Voting Scheme; 2022 - ISSN 1060-992X

11. Szegedy C, Zaremba W, Sutskever I, Bruna J, Erhan D, Goodfellow IJ, Fergus R (2014) Intriguing properties of neural networks. In: Bengio Y, LeCun Y (eds) 2nd international conference on learning representations, ICLR 2014, Banff, AB, Canada, April 14–16, 2014, conference track proceedings, arXiv:1312.6199, DOI: 10.48550/arXiv.1312.6199.

12. Yao Chen1, Yijie Gui1, Hong Lin1, Wensheng Gan1,2∗, Yongdong Wu; Federated Learning Attacks and Defenses: A Survey – 2022; arXiv:2211.14952v1, DOI: 10.48550/arXiv.2211.14952.

13. JI Shou-Ling, Du Tian-Yu, Li Jin-Feng, Shen Chao, Li Bo - Security and privacy of machine learning models: a survey. Ruan Jian Xue Bao/J Softw 32(1):41–67, 2021, DOI: 10.13328/j.cnki.jos.006131.

14. Cohen JM, Rosenfeld E, Kolter JZ - Certified adversarial robustness via randomized smoothing. In: Chaudhuri K, Salakhutdinov R (eds) Proceedings of the 36th international conference on machine learning, ICML 2019, 9–15 June 2019, Long Beach, California, USA, volume 97 of proceedings of machine learning research. PMLR, pp 1310–1320, DOI: 10.48550/arXiv.1902.02918.

15. Nasr M, Shokri R, Houmansadr A (2019) Comprehensive privacy analysis of deep learning: passive and active white-box inference attacks against centralized and federated learning. In: 2019 IEEE symposium on security and privacy (SP). IEEE, pp 739–753, DOI: 10.1109/SP.2019.00065.

16. Ren K, Meng QR, Yan SK - Survey of artificial intelligence data security and privacy protection. Chin J Netw Inf Secur 7(1):1–10, 2021, DOI: 10.11959/j.issn.2096-109x.2021001.

17. Jayaraman B, Evans D (2019) Evaluating differentially private machine learning in practice. In: Heninger N, Traynor P (eds) 28th USENIX security symposium, USENIX security 2019, Santa Clara, CA, USA, August 14–16, 2019. USENIX Association, pp 1895–1912, DOI: 10.48550/arXiv.1902.08874.

18. Melis L, Song C, De Cristofaro E, Shmatikov V (2019) Exploiting unintended feature leakage in collaborative learning. In: 2019 IEEE symposium on security and privacy (SP). IEEE, pp 691–706. DOI: 10.48550/arXiv.1805.04049.

19. Papernot N, McDaniel PD, Sinha A, Wellman MP (2018) Sok: security and privacy in machine learning. In: 2018 IEEE European symposium on security and privacy, EuroS&P 2018, London, United Kingdom, April 24–26, 2018. IEEE, pp 399–414. https://doi.org/10.1109/EuroSP.2018.00035

20. B. Hitaj, G. Ateniese, and F. Perez-Cruz, “Deep models under the gan: information leakage from collaborative deep learning,” in ACM SIGSAC Conference on Computer and Communications Security, 2017, pp. 603–618.

21. L. Zhu, Z. Liu, and S. Han, “Deep leakage from gradients,” Advances in Neural Information Processing Systems, vol. 32, 2019.

22. Geiping J, Bauermeister H, Dröge H, Moeller M (2020) Inverting gradients—how easy is it to break privacy in federated learning? arXiv preprint arXiv:2003.14053, DOI: 10.48550/arXiv.2003.14053.

23. Fang H, Qian Q (2021) Privacy preserving machine learning with homomorphic encryption and federated learning. Future Internet 13(4):94, DOI: 10.3390/fi13040094.

24. Li Y, Zhou Y, Jolfaei A, Dongjin Y, Gaochao X, Zheng X (2020) Privacy-preserving federated learning framework based on chained secure multi-party computing. IEEE Internet Things J. DOI: 10.1109/JIOT.2020.3022911.

25. Lyu L, Yu H, Ma X, Sun L, Zhao J, Yang Q, Yu PS (2020) Privacy and robustness in federated learning: attacks and defenses. arXiv preprint arXiv:2012.06337, DOI: 10.48550/arXiv.2012.06337.