A comparative study of combining deep learning and homomorphic encryption techniques
Keywords:
CNN Deep Learning Homomorphic encryption Privacy preserving
Abstract
Deep learning simulation necessitates a considerable amount of internal computational resources and fast training for large amounts of data. The cloud has been delivering software to help with this transition in recent years, posing additional security risks to data breaches. Modern encryption schemes maintain personal secrecy and are the best method for protecting data stored on a server and data sent from an unauthorized third party. However, when data must be stored or analyzed, decryption is needed, and homomorphic encryption was the first symptom of data security issues found with Strong Encryption.It enables an untrustworthy cloud resource to process encrypted data without revealing sensitive information. This paper looks at the fundamental principles of homomorphic encryption, their forms, and how to integrate them with deep learning. Researchers are particularly interested in privacy-preserving Homomorphic encryption schemes for neural networks. Finally, present options, open problems, threats, prospects, and new research paths are identified across networks
Downloads
Download data is not yet available.
References
[1] X. Jiang, K. Lauter, M. Kim, and Y. Song, “Secure outsourced matrix computation and application to neural networks,” Proc. ACM Conf. Comput. Commun. Secur., pp. 1209–1222, 2018, doi: 10.1145/3243734.3243837.
[2] D. Zhang, X. Chen, D. Wang, and J. Shi, “A survey on collaborative deep learning and privacy-preserving,” Proc. - 2018 IEEE 3rd Int. Conf. Data Sci. Cyberspace, DSC 2018, pp. 652–658, 2018, doi: 10.1109/DSC.2018.00104.
[3] A. Boulemtafes et al., “A review of privacy-preserving techniques for deep learning To cite this version : HAL Id : hal-02921443 A Review of Privacy-Preserving Techniques for Deep Learning,” 2020.
[4] A. QaisarAhmadAlBadawi et al., “Towards the AlexNet Moment for Homomorphic Encryption: HCNN, the First Homomorphic CNN on Encrypted Data with GPUs,” IEEE Trans. Emerg. Top. Comput., 2020, doi: 10.1109/TETC.2020.3014636.
[5] Q. Lou and L. Jiang, “SHE: A fast and accurate deep neural network for encrypted data,” Adv. Neural Inf. Process. Syst., vol. 32, no. NeurIPS, pp. 1–9, 2019.
[6] P. Xie, B. Wu, and G. Sun, “Bayhenn: Combining Bayesian deep learning and homomorphic encryption for secure DNN inference,” IJCAI Int. Jt. Conf. Artif. Intell., vol. 2019-Augus, pp. 4831–4837, 2019, doi: 10.24963/ijcai.2019/671.
[7] J. Chao et al., “CaRENets: Compact and Resource-Efficient CNN for Homomorphic Inference on Encrypted Medical Images,” 2019, [Online]. Available: http://arxiv.org/abs/1901.10074.
[8] S. Obla, X. Gong, A. Aloufi, P. Hu, and D. Takabi, “Effective Activation Functions for Homomorphic Evaluation of Deep Neural Networks,” IEEE Access, vol. 8, pp. 153098–153112, 2020, doi: 10.1109/ACCESS.2020.3017436.
[9] K. Owusu-Agyemeng, Z. Qin, H. Xiong, Y. Liu, T. Zhuang, and Z. Qin, “MSDP: multi-scheme privacy-preserving deep learning via differential privacy,” Pers. Ubiquitous Comput., 2021, doi: 10.1007/s00779-021-01545-0.
[10] J. Suh and T. Tanaka, “Encrypted Value Iteration and Temporal Difference Learning over Leveled Homomorphic Encryption,” Proc. Am. Control Conf., vol. 2021-May, pp. 2555–2561, 2021, doi: 10.23919/ACC50511.2021.9483184.
[11] T. Zhang, Y. Zeng, and B. Xu, “HCNN: A neural network model for combining local and global features towards human-like classification,” Int. J. Pattern Recognit. Artif. Intell., vol. 30, no. 1, 2016, doi: 10.1142/S0218001416550041.
[12] L. Wei, “A Brief Introduction to Deep Learning,” C++ Template Metaprogramming Pract., pp. 69–82, 2020, doi: 10.1201/9781003102311-5.
[13] A. Vizitiu, C. I. Niǎ, A. Puiu, C. Suciu, and L. M. Itu, “Applying Deep Neural Networks over Homomorphic Encrypted Medical Data,” Comput. Math. Methods Med., vol. 2020, 2020, doi: 10.1155/2020/3910250.
[14] M. Hao, H. Li, X. Luo, G. Xu, H. Yang, and S. Liu, “Efficient and Privacy-Enhanced Federated Learning for Industrial Artificial Intelligence,” IEEE Trans. Ind. Informatics, vol. 16, no. 10, pp. 6532–6542, 2020, doi: 10.1109/TII.2019.2945367.
[15] S. M. Ieee, F. Ieee, V. Sze, Y.-H. Chen, T.-J. Yang, and J. S. Emer, “MIT tutorial,” Proc. IEEE, vol. 105, no. 12, pp. 2295–2329, 2017, [Online]. Available: http://ieeexplore.ieee.org/document/8114708/.
[16] B. N. D. Zaid Khalaf Hussien, “Anomaly Detection Approach Based on Deep Neural Network and Dropout,” Baghdad Sci. J., vol. 17, pp. 701–709, 2020.
[17] J. Fan and F. Vercauteren, “Somewhat Practical Fully Homomorphic Encryption,” Proc. 15th Int. Conf. Pract. Theory Public Key Cryptogr., pp. 1–16, 2012, [Online]. Available: https://eprint.iacr.org/2012/144.
[18] B. Macq and Q. Jj, “Cryptography for Trusted Artificial Intelligence in Medicine,” pp. 207–210, 2021, doi: 10.34297/AJBSR.2021.13.001858.R.
[19] L. Lyu, X. He, Y. W. Law, and M. Palaniswami, “Privacy-preserving collaborative deep learning with application to human activity recognition,” Int. Conf. Inf. Knowl. Manag. Proc., vol. Part F1318, no. November, pp. 1219–1228, 2017, doi: 10.1145/3132847.3132990.
[20] M. A. Will and R. K. L. Ko, A guide to homomorphic encryption. Elsevier Inc., 2015.
[21] S. S. Sathya, P. Vepakomma, R. Raskar, R. Ramachandra, and S. Bhattacharya, “A Review of Homomorphic Encryption Libraries for Secure Computation,” pp. 1–12, 2018, [Online]. Available: http://arxiv.org/abs/1812.02428.
[22] B. Pulido-Gaytan et al., “Privacy-preserving neural networks with Homomorphic encryption: Challenges and opportunities,” Peer-to-Peer Netw. Appl., vol. 14, no. 3, pp. 1666–1691, 2021, doi: 10.1007/s12083-021-01076-8.
[23] K. Huang, X. Liu, S. Fu, D. Guo, and M. Xu, “A Lightweight Privacy-Preserving CNN Feature Extraction Framework for Mobile Sensing,” IEEE Trans. Dependable Secur. Comput., vol. 18, no. 3, pp. 1441–1455, 2021, doi: 10.1109/TDSC.2019.2913362.
[24] W. Sirichotedumrong and H. Kiya, “A GAN-based image transformation scheme for privacy-preserving deep neural networks,” Eur. Signal Process. Conf., vol. 2021-Janua, no. c, pp. 745–749, 2021, doi: 10.23919/Eusipco47968.2020.9287532.
[2] D. Zhang, X. Chen, D. Wang, and J. Shi, “A survey on collaborative deep learning and privacy-preserving,” Proc. - 2018 IEEE 3rd Int. Conf. Data Sci. Cyberspace, DSC 2018, pp. 652–658, 2018, doi: 10.1109/DSC.2018.00104.
[3] A. Boulemtafes et al., “A review of privacy-preserving techniques for deep learning To cite this version : HAL Id : hal-02921443 A Review of Privacy-Preserving Techniques for Deep Learning,” 2020.
[4] A. QaisarAhmadAlBadawi et al., “Towards the AlexNet Moment for Homomorphic Encryption: HCNN, the First Homomorphic CNN on Encrypted Data with GPUs,” IEEE Trans. Emerg. Top. Comput., 2020, doi: 10.1109/TETC.2020.3014636.
[5] Q. Lou and L. Jiang, “SHE: A fast and accurate deep neural network for encrypted data,” Adv. Neural Inf. Process. Syst., vol. 32, no. NeurIPS, pp. 1–9, 2019.
[6] P. Xie, B. Wu, and G. Sun, “Bayhenn: Combining Bayesian deep learning and homomorphic encryption for secure DNN inference,” IJCAI Int. Jt. Conf. Artif. Intell., vol. 2019-Augus, pp. 4831–4837, 2019, doi: 10.24963/ijcai.2019/671.
[7] J. Chao et al., “CaRENets: Compact and Resource-Efficient CNN for Homomorphic Inference on Encrypted Medical Images,” 2019, [Online]. Available: http://arxiv.org/abs/1901.10074.
[8] S. Obla, X. Gong, A. Aloufi, P. Hu, and D. Takabi, “Effective Activation Functions for Homomorphic Evaluation of Deep Neural Networks,” IEEE Access, vol. 8, pp. 153098–153112, 2020, doi: 10.1109/ACCESS.2020.3017436.
[9] K. Owusu-Agyemeng, Z. Qin, H. Xiong, Y. Liu, T. Zhuang, and Z. Qin, “MSDP: multi-scheme privacy-preserving deep learning via differential privacy,” Pers. Ubiquitous Comput., 2021, doi: 10.1007/s00779-021-01545-0.
[10] J. Suh and T. Tanaka, “Encrypted Value Iteration and Temporal Difference Learning over Leveled Homomorphic Encryption,” Proc. Am. Control Conf., vol. 2021-May, pp. 2555–2561, 2021, doi: 10.23919/ACC50511.2021.9483184.
[11] T. Zhang, Y. Zeng, and B. Xu, “HCNN: A neural network model for combining local and global features towards human-like classification,” Int. J. Pattern Recognit. Artif. Intell., vol. 30, no. 1, 2016, doi: 10.1142/S0218001416550041.
[12] L. Wei, “A Brief Introduction to Deep Learning,” C++ Template Metaprogramming Pract., pp. 69–82, 2020, doi: 10.1201/9781003102311-5.
[13] A. Vizitiu, C. I. Niǎ, A. Puiu, C. Suciu, and L. M. Itu, “Applying Deep Neural Networks over Homomorphic Encrypted Medical Data,” Comput. Math. Methods Med., vol. 2020, 2020, doi: 10.1155/2020/3910250.
[14] M. Hao, H. Li, X. Luo, G. Xu, H. Yang, and S. Liu, “Efficient and Privacy-Enhanced Federated Learning for Industrial Artificial Intelligence,” IEEE Trans. Ind. Informatics, vol. 16, no. 10, pp. 6532–6542, 2020, doi: 10.1109/TII.2019.2945367.
[15] S. M. Ieee, F. Ieee, V. Sze, Y.-H. Chen, T.-J. Yang, and J. S. Emer, “MIT tutorial,” Proc. IEEE, vol. 105, no. 12, pp. 2295–2329, 2017, [Online]. Available: http://ieeexplore.ieee.org/document/8114708/.
[16] B. N. D. Zaid Khalaf Hussien, “Anomaly Detection Approach Based on Deep Neural Network and Dropout,” Baghdad Sci. J., vol. 17, pp. 701–709, 2020.
[17] J. Fan and F. Vercauteren, “Somewhat Practical Fully Homomorphic Encryption,” Proc. 15th Int. Conf. Pract. Theory Public Key Cryptogr., pp. 1–16, 2012, [Online]. Available: https://eprint.iacr.org/2012/144.
[18] B. Macq and Q. Jj, “Cryptography for Trusted Artificial Intelligence in Medicine,” pp. 207–210, 2021, doi: 10.34297/AJBSR.2021.13.001858.R.
[19] L. Lyu, X. He, Y. W. Law, and M. Palaniswami, “Privacy-preserving collaborative deep learning with application to human activity recognition,” Int. Conf. Inf. Knowl. Manag. Proc., vol. Part F1318, no. November, pp. 1219–1228, 2017, doi: 10.1145/3132847.3132990.
[20] M. A. Will and R. K. L. Ko, A guide to homomorphic encryption. Elsevier Inc., 2015.
[21] S. S. Sathya, P. Vepakomma, R. Raskar, R. Ramachandra, and S. Bhattacharya, “A Review of Homomorphic Encryption Libraries for Secure Computation,” pp. 1–12, 2018, [Online]. Available: http://arxiv.org/abs/1812.02428.
[22] B. Pulido-Gaytan et al., “Privacy-preserving neural networks with Homomorphic encryption: Challenges and opportunities,” Peer-to-Peer Netw. Appl., vol. 14, no. 3, pp. 1666–1691, 2021, doi: 10.1007/s12083-021-01076-8.
[23] K. Huang, X. Liu, S. Fu, D. Guo, and M. Xu, “A Lightweight Privacy-Preserving CNN Feature Extraction Framework for Mobile Sensing,” IEEE Trans. Dependable Secur. Comput., vol. 18, no. 3, pp. 1441–1455, 2021, doi: 10.1109/TDSC.2019.2913362.
[24] W. Sirichotedumrong and H. Kiya, “A GAN-based image transformation scheme for privacy-preserving deep neural networks,” Eur. Signal Process. Conf., vol. 2021-Janua, no. c, pp. 745–749, 2021, doi: 10.23919/Eusipco47968.2020.9287532.
Published
2022-03-13
How to Cite
Alsaedi, E., & kadhim, A. (2022). A comparative study of combining deep learning and homomorphic encryption techniques. Al-Qadisiyah Journal of Pure Science, 27(1), comp 17-33. https://doi.org/10.29350/qjps.2022.27.1.1452
Issue
Section
Computer
Copyright © Author(s) . This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.
