現今有許多系統服務會使用深度學習技術，且可能會使用到敏感資料，所以須在隱私保護的情況下進行計算。安全多方運算則是其中一個方法，而安全多方運算中的安全雙方運算，很適合使用在深度學習的服務中，可以保護使用者資料及模型參數。在安全雙方運算中，有大量時間會用來進行與使用者輸入資料無關的運算，如計算加密資源。為了減少使用者等待時間，我們提出了預處理運算的方法。將這些與輸入資料無關的運算事先算完，並存起來。當使用者需要服務，提供資料後直接使用預處理的資源進行安全雙方運算，便可以大幅減少時間。在微基準測試及卷積神經網路的實驗中，速度分別提升了6倍及4倍。

While deep learning techniques have been widely used in many application domains, security issues are raised when dealing with sensitive data.
Secure computation is made possible by privacy-preserving computing without leaking the content of the data being used. One of the potential methods for privacy-preserving computing is secure multi party computation (MPC). A variation of the secure multi-party computation is the secure two-party computation (TPC). TPC is appropriate for a deep learning application's secure model inference, which can aid in safeguarding model parameters and user input data. During TPC execution, the majority of the time is spent in doing input-invariant computations, such as using cryptographic resources. The proposed precomputing scheme accelerates the time of the TPC execution by performing the input-independent computations before the TPC execution. Our experimental results show up to 6x and 3.2x performance speedups than the state-of-the-art work, TONIC, for the microbenchmark program and for the deep learning model, MobileNetV2, respectively.

[1] Abbas Acar, Hidayet Aksu, A Selcuk Uluagac, and Mauro Conti. A survey on homomorphic encryption schemes: Theory and implementation. ACM Computing Surveys (CSUR), 51(4):1–35, 2018.
[2] Gilad Asharov, Yehuda Lindell, Thomas Schneider, and Michael Zohner. More efficient oblivious transfer and extensions for faster secure computation. In Proceedings of the 2013 ACM SIGSAC conference on Computer & communications security, pages 535–548. ACM, 2013.
[3] Donald Beaver. Efficient multiparty protocols using circuit randomization. In Annual International Cryptology Conference, volume 576, pages 420–432. Springer, 1991.
[4] Aner Ben-Efraim, Michael Nielsen, and Eran Omri. Turbospeedz: Double your online spdz! improving spdz using function dependent preprocessing. In International Conference on Applied Cryptography and Network Security, pages 530–549. Springer, 2019.
[5] Fabian Boemer, Anamaria Costache, Rosario Cammarota, and Casimir Wierzynski. ngraph-he2: A high-throughput framework for neural network inference on encrypted data. In ACM Workshop on Encrypted Computing & Applied Homomorphic Cryptography, pages 45–56. ACM, 2019.
[6] Fabian Boemer, Yixing Lao, Rosario Cammarota, and Casimir Wierzynski. ngraphhe: a graph compiler for deep learning on homomorphically encrypted data. In ACM International Conference on Computing Frontiers, pages 3–13. ACM, 2019.
[7] Lennart Braun, Daniel Demmler, Thomas Schneider, and Oleksandr Tkachenko. Motion– a framework for mixed-protocol multi-party computation. ACM Transactions on Privacy and Security, 25(2), 2022.
[8] Niklas Büscher and Stefan Katzenbeisser. Compilation for Secure Multi-party Computation. Springer Briefs in Computer Science. Springer, 2017.
[9] Ivan Damgård, Valerio Pastro, Nigel Smart, and Sarah Zakarias. Multiparty computation from somewhat homomorphic encryption. In Annual Cryptology Conference, pages 643–662. Springer, 2012.
[10] Paolo D’Arco and Roberto De Prisco. Secure two-party computation: A visual way. In Carles Padró, editor, Information Theoretic Security - 7th International Conference, ICITS 2013, Singapore, November 28-30, 2013, Proceedings, volume 8317 of Lecture Notes in Computer Science, pages 18–38. Springer, 2013.
[11] Daniel Demmler, Thomas Schneider, and Michael Zohner. Aby - a framework for efficient mixed-protocol secure two-party computation. In Network and Distributed System Security Symposium. The Internet Society, 2015.
[12] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In IEEE conference on computer vision and pattern recognition, pages 248–255. IEEE Computer Society, 2009.
[13] Li Deng. The mnist database of handwritten digit images for machine learning research. IEEE Signal Processing Magazine, 29(6):141–142, 2012.
[14] David Evans, Vladimir Kolesnikov, and Mike Rosulek. A pragmatic introduction to secure multi-party computation. Found. Trends Priv. Secur., 2(2-3):70–246, 2018.
[15] Oded Goldreich, Silvio Micali, and Avi Wigderson. How to play any mental game or a completeness theorem for protocols with honest majority. In ACM Symposium on Theory of Computing, pages 218–229. ACM, 1987.
[16] Po-Hsuan Huang, Chia-Heng Tu, and Shen-Ming Chung. Tonic: Towards oblivious neural inference compiler. In Proceedings of the 36th Annual ACM Symposium on Applied Computing, pages 491–500. ACM, 2021.
[17] Marcel Keller. Mp-spdz: A versatile framework for multi-party computation. In Proceedings of the 2020 ACM SIGSAC Conference on Computer and Communications Security, pages 1575–1590. ACM, 2020.
[18] Vladimir Kolesnikov and Thomas Schneider. Improved garbled circuit: Free xor gates and applications. In International Colloquium on Automata, Languages, and Programming, pages 486–498. Springer, 2008.
[19] Chang Liu, Xiao Shaun Wang, Kartik Nayak, Yan Huang, and Elaine Shi. Oblivm: A programming framework for secure computation. In IEEE Symposium on Security and Privacy, pages 359–376. IEEE Computer Society, 2015.
[20] Arpita Patra, Thomas Schneider, Ajith Suresh, and Hossein Yalame. Aby2.0: Improved mixed-protocol secure two-party computation. In USENIX Security Symposium, pages 2165–2182. USENIX Association, 2021.
[21] Deevashwer Rathee, Mayank Rathee, Nishant Kumar, Nishanth Chandran, Divya Gupta, Aseem Rastogi, and Rahul Sharma. Cryptflow2: Practical 2-party secure inference. In ACM SIGSAC Conference on Computer and Communications Security, pages 325–342. ACM, 2020.
[22] Bita Darvish Rouhani, M. Sadegh Riazi, and Farinaz Koushanfar. Deepsecure: scalable provably-secure deep learning. In Design Automation Conference, pages 1–6. ACM, 2018.
[23] Mark Sandler, Andrew G. Howard, Menglong Zhu, Andrey Zhmoginov, and Liang-Chieh Chen. Mobilenetv2: Inverted residuals and linear bottlenecks. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4510–4520. Computer Vision Foundation / IEEE Computer Society, 2018.
[24] Xiao Wang, Alex J. Malozemoff, and Jonathan Katz. EMP-toolkit: Efficient MultiParty computation toolkit. https://github.com/emp-toolkit, 2016.
[25] Andrew Chi-Chih Yao. Protocols for secure computations. In IEEE Symposium on Foundations of Computer Science, pages 160–164. IEEE Computer Society, 1982.
[26] Samee Zahur and David Evans. Obliv-c: A language for extensible data-oblivious computation. IACR Cryptology ePrint Archive, 2015:1153, 2015.
Part of this work is published in The Journal of Supercomputing and the information of the journal article is as follows. Huang, PH., Chang, TW., Tu, CH. et al. POPS: an off-peak precomputing scheme for privacy-preserving computing. J Supercomput (2022). https://doi.org/10.1007/s11227-022-04552-x