Abstract
Machine learning applications are spreading in many fields and more often than not manipulate private data in order to derive classifications impacting the lives of many individuals. In this context, it becomes important to work on privacy preserving mechanisms associated to different privacy scenarios: protecting the training data, the classification data, the weights of a neural network. In this paper, we study the possibility of using FHE techniques to address the above issues. In particular, we are able to evaluate a neural network where both its topology and its weights as well as the user data it operates on remain sealed in the encrypted domain. We do so by relying on Hopfield neural networks which are much more “FHE friendly” than their feed-forward counterparts. In doing so, we thus also argue the case of considering different (yet existing) Neural Network models better adapted to FHE, in order to more efficiently address real-world applications.The paper is concluded by experimental results on a face recognition application demonstrating the ability of the approach to provide reasonable recognition timings (\({\approx }0.6\) s) on a single standard processor core.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Albrecht, M.R., Player, R., Scott, S.: On the concrete hardness of learning with errors. Cryptology ePrint Archive, Report 2015/046 (2015)
Bansal, A., Chen, T., Zhong, S.: Privacy preserving back-propagation neural network learning over arbitrarily partitioned data. Neural Comput. Appl. 20, 143–150 (2011)
Bourse, F., Minelli, M., Minihold, M., Paillier, P.: Fast homomorphic evaluation of deep discretized neural networks. In: Shacham, H., Boldyreva, A. (eds.) CRYPTO 2018. LNCS, vol. 10993, pp. 483–512. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-96878-0_17
Chabanne, H., de Wargny, A., Milgram, J., Morel, C., Prouff, E.: Privacy-preserving classification on deep neural network. IACR Cryptology ePrint Archive 2017 (2017)
Chillotti, I., Gama, N., Georgieva, M., Izabachène, M.: Faster fully homomorphic encryption: bootstrapping in less than 0.1 seconds. In: Cheon, J.H., Takagi, T. (eds.) ASIACRYPT 2016. LNCS, vol. 10031, pp. 3–33. Springer, Heidelberg (2016). https://doi.org/10.1007/978-3-662-53887-6_1
Chillotti, I., Gama, N., Georgieva, M., Izabachène, M.: Improving TFHE: faster packed homomorphic operations and efficient circuit bootstrapping. Cryptology ePrint Archive, Report 2017/430 (2017). https://eprint.iacr.org/2017/430
Gentry, C., Sahai, A., Waters, B.: Homomorphic encryption from learning with errors: conceptually-simpler, asymptotically-faster, attribute-based. In: Canetti, R., Garay, J.A. (eds.) CRYPTO 2013. LNCS, vol. 8042, pp. 75–92. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-40041-4_5
Gurney, K.: An introduction to neural networks (1997)
Hare, J.S., Samangooei, S., Dupplaw, D.P.: OpenIMAJ and ImageTerrier: Java libraries and tools for scalable multimedia analysis and indexing of images. In: Proceedings of the 19th ACM International Conference on Multimedia. ACM (2011)
Hopfield, J.J.: Neural networks and physical systems with emergent collective computational abilities. Proc. Natl. Acad. Sci. 79, 2554–2558 (1982)
Lyubashevsky, V., Peikert, C., Regev, O.: On ideal lattices and learning with errors over rings. In: Gilbert, H. (ed.) EUROCRYPT 2010. LNCS, vol. 6110, pp. 1–23. Springer, Heidelberg (2010). https://doi.org/10.1007/978-3-642-13190-5_1
MacKay, D.: Information theory, inference, and learning algorithms (2003)
Shokri, R., Shmatikov, V.: Privacy-preserving deep learning. In: 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton) (2015)
Sulehria, H.K., Zhang, Y.: Hopfield neural networks: a survey. In: Proceedings of the 6th WSEAS Conference on Artificial Intelligence, Knowledge Engineering and Data Bases. WSEAS (2007)
Tan, X., Triggs, B.: Enhanced local texture feature sets for face recognition under difficult lighting conditions. Trans. Image Process. 19, 1635–1650 (2010)
Xie, P., Bilenko, M., Finley, T., Gilad-Bachrach, R., Lauter, K.E., Naehrig, M.: Crypto-Nets: neural networks over encrypted data. CoRR (2014)
Yuan, J., Yu, S.: Privacy preserving back-propagation neural network learning made practical with cloud computing. IEEE Trans. Parallel Distrib. Syst. 25, 212–221 (2014)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
Izabachène, M., Sirdey, R., Zuber, M. (2019). Practical Fully Homomorphic Encryption for Fully Masked Neural Networks. In: Mu, Y., Deng, R., Huang, X. (eds) Cryptology and Network Security. CANS 2019. Lecture Notes in Computer Science(), vol 11829. Springer, Cham. https://doi.org/10.1007/978-3-030-31578-8_2
Download citation
DOI: https://doi.org/10.1007/978-3-030-31578-8_2
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-31577-1
Online ISBN: 978-3-030-31578-8
eBook Packages: Computer ScienceComputer Science (R0)