Abstract
Convolutional Neural networks have been widely adopted in computer vision because of their robust performance in a variety of applications. Many of the image detection and classification algorithms are being offered as a service by the Cloud service providers. In its current manifestation, the service provider has access to the rich data that is sent as a query thereby compromising the privacy of the user. Encryption can be used to ensure security during transmission. However, it cannot ensure data privacy. In order to protect the privacy and ensure security of the user data, there is a need to develop new approaches that can perform computer vision tasks on encrypted data. In this paper, Fully Homomorphic Encryption (FHE) is used to ensure security and privacy of the data. The proposed method builds the necessary algorithms to allow the server to make inferences on the encrypted input and give encrypted result back to the user. Deep neural network for FHE encrypted data is computationally very heavy. In order to address this problem, variable length packing on the pruned deep learning model is employed. An algorithm to automatically recommend appropriate parameters for pruning and variable length packing is proposed. Further, user’s preference for Quality of Service (QoS) is taken into account in the proposed framework. CIFAR-10 image dataset is used to evaluate the method in terms of performance and accuracy for a ten class classification. Our experimental analysis show upto 60% improvement in terms of performance while using our optimizations.
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
Ahn, B.H., Pilligundla, A., Yazdanbakhsh, A., Esmaeilzadeh, H.: Chameleon: adaptive code optimization for expedited deep neural network compilation (2020)
Albrecht, M.: LWE-estimator. https://bitbucket.org/malb/lwe-estimator/src/master. Accessed 14 Apr 2021
Al Badawi, A., et al.: The alexnet moment for homomorphic encryption: HCNN, the first homomorphic CNN on encrypted data with GPUs. CoRR abs/1811.00778 (2018)
Balaei, S.: Decision radar ez-topsis. https://decision-radar.com/Topsis.html
Chabanne, H., de Wargny, A., Milgram, J., Morel, C., Prouff, E.: Privacy-preserving classification on deep neural network. IACR Cryptol. ePrint Arch. 2017, 35 (2017)
Cheon, J.H., Kim, A., Kim, M., Song, Y.: Homomorphic encryption for arithmetic of approximate numbers. In: Takagi, T., Peyrin, T. (eds.) ASIACRYPT 2017. LNCS, vol. 10624, pp. 409–437. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-70694-8_15
Chou, E., Beal, J., Levy, D., Yeung, S., Haque, A., Fei-Fei, L.: Faster cryptonets: leveraging sparsity for real-world encrypted inference (2018)
Dowlin, N., Gilad-Bachrach, R., Laine, K., Lauter, K., Naehrig, M., Wernsing, J.: Cryptonets: applying neural networks to encrypted data with high throughput and accuracy. In: Proceedings of the 33rd ICML - Volume 48, pp. 201–210. JMLR.org (2016)
Gentry, C.: A fully homomorphic encryption scheme. Ph.D. thesis, Stanford University (2009). crypto.stanford.edu/craig
Goodfellow, I., Bengio, Y., Courville, A., Bengio, Y.: Deep Learning, vol. 1. MIT Press, Cambridge (2016)
Hesamifard, E., Takabi, H., Ghasemi, M.: Cryptodl: deep neural networks over encrypted data. CoRR abs/1711.05189 (2017)
Hunt, T., Song, C., Shokri, R., Shmatikov, V., Witchel, E.: Chiron: privacy-preserving machine learning as a service. arXiv preprint arXiv:1803.05961 (2018)
Intel: Intel’s threading building blocks. https://software.intel.com/content/www/us/en/develop/tools/oneapi/components/onetbb.html
Ishiyama, T., Suzuki, T., Yamana, H.: Highly accurate CNN inference using approximate activation functions over homomorphic encryption (2020)
Juvekar, C., Vaikuntanathan, V., Chandrakasan, A.: Gazelle: a low latency framework for secure neural network inference. CoRR abs/1801.05507 (2018)
Kim, M., Jiang, X., Lauter, K., Ismayilzada, E., Shams, S.: HEAR: human action recognition via neural networks on homomorphically encrypted data (2021)
Liu, J., Juuti, M., Lu, Y., Asokan, N.: Oblivious neural network predictions via minionn transformations. In: Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security, pp. 619–631. ACM, New York, NY, USA (2017)
Mohassel, P., Zhang, Y.: Secureml: a system for scalable privacy-preserving machine learning. In: 2017 IEEE Symposium on Security and Privacy (SP), pp. 19–38. IEEE (2017)
Nandakumar, K., Ratha, N., Pankanti, S., Halevi, S.: Towards deep neural network training on encrypted data. In: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp. 40–48 (2019)
Riazi, M.S., Samragh, M., Chen, H., Laine, K., Lauter, K., Koushanfar, F.: XONN: Xnor-based oblivious deep neural network inference. In: 28th USENIX Security Symposium (USENIX Security 19), pp. 1501–1518 (2019)
Rouhani, B.D., Riazi, M.S., Koushanfar, F.: DeepSecure: scalable provably-secure deep learning. In: Proceedings of the 55th Annual Design Automation Conference, pp. 1–6 (2018)
Sanyal, A., Kusner, M.J., Gascon, A., Kanade, V.: TAPAS: tricks to accelerate (encrypted) prediction as a service. arXiv preprint arXiv:1806.03461 (2018)
Microsoft SEAL (release 3.1). Microsoft Research, Redmond, WA, December 2018. https://github.com/Microsoft/SEAL
Shoup, V.: Number theory library. https://libntl.org/doc/tour.html
Valerio, L., Nardini, F.M., Passarella, A., Perego, R.: Dynamic hard pruning of neural networks at the edge of the internet (2020)
Wagh, S., Gupta, D., Chandran, N.: SecureNN: efficient and private neural network training. IACR Cryptol. ePrint Arch. 2018, 442 (2018)
Xu, R., Joshi, J.B.D., Li, C.: CryptoNN: training neural networks over encrypted data. CoRR abs/1904.07303 (2019)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Shaik, I., Chaudhari, R., Rajan, M.A., Gubbi, J., Balamuralidhar, P., Lodha, S. (2021). WiP: QoS Based Recommendation System for Efficient Private Inference of CNN Using FHE. In: Tripathy, S., Shyamasundar, R.K., Ranjan, R. (eds) Information Systems Security. ICISS 2021. Lecture Notes in Computer Science(), vol 13146. Springer, Cham. https://doi.org/10.1007/978-3-030-92571-0_13
Download citation
DOI: https://doi.org/10.1007/978-3-030-92571-0_13
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-92570-3
Online ISBN: 978-3-030-92571-0
eBook Packages: Computer ScienceComputer Science (R0)