Abstract
Adopting mobile healthcare network (MHN) services such as disease detection is fraught with concerns about the security and privacy of the entities involved and the resource restrictions at the Internet of Things (IoT) nodes. Hence, the essential requirements for disease detection services are to (i) produce accurate and fast disease detection without jeopardizing the privacy of health clouds and medical users and (ii) reduce the computational and transmission overhead (energy consumption) of the IoT devices while maintaining the privacy. For privacy preservation of widely used neural network– (NN) based disease detection, existing literature suggests either computationally heavy public key fully homomorphic encryption (FHE), or secure multiparty computation, with a large number of interactions. Hence, the existing privacy-preserving NN schemes are energy consuming and not suitable for resource-constrained IoT nodes in MHN. This work proposes a lightweight, fully homomorphic, symmetric key FHE scheme (SkFhe) to address the issues involved in implementing privacy-preserving NN. Based on SkFhe, widely used non-linear activation functions ReLU and Leaky ReLU are implemented over the encrypted domain. Furthermore, based on the proposed privacy-preserving linear transformation and non-linear activation functions, an energy-efficient, accurate, and privacy-preserving NN is proposed. The proposed scheme guarantees privacy preservation of the health cloud’s NN model and medical user’s data. The experimental analysis demonstrates that the proposed solution dramatically reduces the overhead in communication and computation at the user side compared to the existing schemes. Moreover, the improved energy efficiency at the user is accomplished with reduced diagnosis time without sacrificing classification accuracy.
- [1] . 2015. Security and privacy for mobile healthcare networks: From a quality of protection perspective. IEEE Wireless Commun. 22, 4 (August 2015), 104–112.
DOI: Google ScholarDigital Library - [2] . 2018. Privacy preservation in e-healthcare environments: State of the art and future directions. IEEE Access 6, (Februarty 2018), 464–478.
DOI: Google ScholarCross Ref - [3] . 2021. Security and privacy requirements for the internet of things: A survey. ACM Trans. Internet Things 2, 1, Article 6 (February 2021), 1–37.
DOI: Google ScholarDigital Library - [4] . 2012. Artificial neural network in diagnosis of metastatic carcinoma in effusion cytology. Cytom. Part B Clin. Cytom. 82B, 2 (March 2012), 107–111.
DOI: Google ScholarCross Ref - [5] . 2012. A new approach for locating the minor apical foramen using an artificial neural network. Int. Endod. J. 45, 3 (March 2012), 257–265.
DOI: Google ScholarCross Ref - [6] . 2019. Privacy-preserving deep learning via weight transmission. IEEE Trans. Inf. Forens. Secur. 14, 11 (November 2019), 3003–3015.
DOI: Google ScholarDigital Library - [7] . 2020. The role of neural network activation functions. IEEE Sign. Process Lett. 27, (September 2020), 1779–1783.
DOI: Google ScholarCross Ref - [8] . 2018. vReLU Activation functions for artificial neural networks. In Proceedings of the 14th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD’18), 856–860.
DOI: Google ScholarCross Ref - [9] . 2017. Activation functions of deep neural networks for polar decoding applications. In Proceedings of the IEEE 28th Annual International Symposium on Personal, Indoor, and Mobile Radio Communications (PIMRC’17). 1–5.
DOI: Google ScholarDigital Library - [10] . 2017. Oblivious neural network ppredictions via MiniONN transformations. In Proceedings of the ACM SIGSAC Conference on Computer and Communications Security. 619–631.Google Scholar
- [11] . 2020. Lightweight privacy-preserving training and evaluation for discretized neural networks. IEEE IoT J. 7, 4 (April 2020), 2663–2678.
DOI: Google ScholarCross Ref - [12] . 2019. MSCryptoNet: Multi-scheme privacy-preserving deep learning in cloud computing. IEEE Access 7, (February 2019), 29344–29354.
DOI: Google ScholarCross Ref - [13] . 2019. Homo-ELM: fully homomorphic extreme learning machine. Int. J. Mach. Learn. Cyber. 11, (December 2019), 1531–1540.
DOI: Google ScholarCross Ref - [14] . 2016. Privacy-preserving logistic regression with distributed data sources via homomorphic encryption, IEICE Trans. Inf. Syst. 99-D, 8 (August 2016), 2079–2089.
DOI: Google ScholarCross Ref - [15] . 2016. Scalable and secure logistic regression via homomorphic encryption. In Proceedings of the 6th ACM Conference on Data and Application Security and Privacy. 142–144.
DOI: Google ScholarDigital Library - [16] . 2017. A survey on security and privacy issues in Internet-of-Things. IEEE IoT J. 4, 5 (October 2017), 1250–1258.
DOI: Google ScholarCross Ref - [17] . 2015. Low-energy security: Limits and opportunities in the Internet of Things. IEEE Secur. Priv. 13, 1 (February 2015), 14–21.
DOI: Google ScholarDigital Library - [18] . 2015. PPDM: A privacy-preserving protocol for cloud-assisted e-healthcare systems. IEEE J. Sel. Top. Sign. Process. 9, 7 (October 2015), 1332–1344.
DOI: Google ScholarCross Ref - [19] . 1978. On data banks and privacy homomorphisms. In Foundations of Secure Computation. Academic Press, New York, NY, 169–179.Google Scholar
- [20] . 2022. Private and energy-efficient decision tree-based disease detection for resource-constrained medical users in mobile healthcare network, IEEE Access 10, (February 2022), 17098–17112.
DOI: Google ScholarCross Ref - [21] . 2017. Linear complexity of second order PN sequences addition with single order PN sequence in nonlinear filter generator, J. Discr. Math. Sci. Cryptogr. 20, 5 (November 2017), 1173–1181.
DOI: Google ScholarCross Ref - [22] . 2022. Lightweight privacy-preserving medical diagnosis in edge computing, IEEE Trans. Serv. Comput. 15, 3 (June 2022), 1606–1618.
DOI: Google ScholarCross Ref - [23] . 2021. Efficient on-site confirmatory testing for atrial fibrillation with derived 12-lead ECG in a wireless body area network. J Amb. Intell. Hum. Comput. (November 2021), 1–19.
DOI: Google ScholarCross Ref - [24] . 2016. CryptoNets: Applying neural networks to encrypted data with high throughput and accuracy. In Proceedings of the 33rd International Conference on International Conference on Machine Learning. 201–210.Google Scholar
- [25] . 2000. Generalized inversion attack on nonlinear filter generators, IEEE Trans. Comput. 49, 10 (October 2000), 1100–1109.
DOI: Google ScholarDigital Library - [26] . 2012. (Leveled) Fully homomorphic encryption without bootstrapping. In Proceedings of the 3rd Innovations in Theoretical Computer Science Conference. 309–325.
DOI: Google ScholarDigital Library - [27] . 2022. Privacy-Preserving machine learning with fully homomorphic encryption for deep neural network. IEEE Access 10, (March 2022), 30039–30054.
DOI: Google ScholarCross Ref - [28] . 2006. A privacy-preserving protocol for neural network-based computation. In Proceedings of the 8th Workshop on Multimedia and Security (MM&Sec’06). ACM, New York, NY, 146–151.
DOI: Google ScholarDigital Library - [29] . 1979. How to share a secret, Commun. ACM 22, 11 (November 1979), 612–613.
DOI: Google ScholarDigital Library - [30] . 2009. Symmetric-key homomorphic encryption for encrypted data processing. In Proceedings of the IEEE International Conference on Communications. 1–5.
DOI: Google ScholarCross Ref - [31] . Recommendation for Key Management: Part 1–General. Retrieved September 13, 2022 from https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-57pt1r5.pdf.Google Scholar
- [32] . How Many Primes Are There? Retrieved June 10, 2021 from https://primes.utm.edu/howmany.html#pnt.Google Scholar
- [33] . Retrieved June 10, 2022 from https://www.cs.utexas.edu/dwu4/notes/CS255LectureNotes.pdf.Google Scholar
- [34] . 2020. Introduction to Modern Cryptography, Vol. 3. CRC Press, Boca Raton, FL.
DOI: Google ScholarCross Ref - [35] . 2009. Foundations of Cryptography: Volume 2, Basic Applications, Vol. 2. Cambridge University Press, Cambridge, UK.
DOI: Google ScholarCross Ref - [36] . 2009. A proof of security of YAOS protocol for two-party computation. J. Cryptol 22, 2 (April 2009), 161–188.
DOI: Google ScholarDigital Library - [37] . 2016. Privacy-preserving patient-centric clinical decision support system on naïve Bayesian classification. IEEE J. Biomed. Health Inform. 20, 2 (March 2016), 655–668.
DOI: Google ScholarCross Ref - [38] . 2016. An efficient privacy-preserving outsourced calculation toolkit with multiple keys. IEEE Trans. Inf. Forens. Secur. 11, 11 (November 2016), 2401–2414.
DOI: Google ScholarDigital Library - [39] . 2018. Automatic Identification of the Rhythm/Morphology Abnormalities in 12-Lead ECGs. Retrieved June 10, 2021 from http://2018.icbeb.org/Challenge.html.Google Scholar
- [40] . 2018. Private machine learning classification based on fully homomorphic encryption. IEEE T. Emerg. Top. Com. 8, 2 (April/June 2018), 352–364.
DOI: Google ScholarCross Ref - [41] . 2013. Design and Implementation of a Homomorphic-Encryption Library. Retrieved from https://github.com/shaih/HElib.Google Scholar
- [42] . 2021. Privacy-preserving outsourced clinical decision support system in the cloud, IEEE Trans. Serv. Comput. 14, 1 (January 2021), 222–234.
DOI: Google ScholarCross Ref - [43] . 2013. Improved security for a ring-based fully homomorphic encryption scheme. In Proceedings of the IMA International Conference on Cryptography and Coding (IMACC’13), Lecture Notes in Computer Science, M. Stam (Ed.), 45–64.
DOI: Google ScholarCross Ref - [44] . 2019. Privacy-preserving and high-accurate outsourced disease predictor on Random Forest. Inf. Sci. 496, (September 2019), 225–241.
DOI: Google ScholarDigital Library - [45] . 1999. Public-key cryptosystems based on composite degree residuosity classes. In Proceedings of the Annual International Conference on the Theory and Applications of Cryptology and Information Security (EUROCRYPT’99), Lecture Notes in Computer Science, Vol. 1592, J. Stern (Ed.). Springer, 223238.
DOI: Google ScholarCross Ref - [46] . 2014. A comparison of the homomorphic encryption schemes fv and yashe. In Proceedings of the Annual International Conference on the Theory and Applications of Cryptology (AFRICACRYPT’14). Springer, Cham, 318-335.
DOI: Google ScholarCross Ref - [47] . 2021. Enabling privacy-assured fog-based data aggregation in E-healthcare systems. IEEE Trans. Industr. Inform. 17, 3 (March 2021), 1948–1957.
DOI: Google ScholarCross Ref - [48] . 2021. Collusion resistant secret sharing scheme for secure data storage and processing over cloud. J. Inf. Secur. Appl. 60, (August 2021).
DOI: Google ScholarCross Ref - [49] . 2018. Energy-aware bio-signal compressed sensing reconstruction on the WBSN-gateway. IEEE Trans. Emerg. Top. Comput. 6, 3 (July/September 2018), 370-381.
DOI: Google ScholarCross Ref
Index Terms
- Energy Efficient and Secure Neural Network–based Disease Detection Framework for Mobile Healthcare Network
Recommendations
An efficient and secure data sharing framework using homomorphic encryption in the cloud
Cloud-I '12: Proceedings of the 1st International Workshop on Cloud IntelligenceDue to cost-efficiency and less hands-on management, data owners are outsourcing their data to the cloud which can provide access to the data as a service. However, by outsourcing their data to the cloud, the data owners lose control over their data as ...
Efficient Integer Vector Homomorphic Encryption Using Deep Learning for Neural Networks
Neural Information ProcessingAbstractMachine learning techniques based on neural networks have achieved significant applications in a wide variety of areas. There is a great risk on disclosing users’ privacy when we train a high-performance model with a large number of datasets ...
Towards Practical Secure Neural Network Inference: The Journey So Far and the Road Ahead
Neural networks (NNs) have become one of the most important tools for artificial intelligence. Well-designed and trained NNs can perform inference (e.g., make decisions or predictions) on unseen inputs with high accuracy. Using NNs often involves ...
Comments