Deep Learning Data Privacy Protection Based on Homomorphic Encryption in AIoT

Abstract

With the rapid development of 5G technology, its high bandwidth, high reliability, low delay, and large connection characteristics have opened up a broader application field of IoT. Moreover, AIoT (Artificial Intelligence Internet of Things) has become the new development direction of IoT. Through deep learning of real-time data provided by the Internet of Things, AI can judge user habits more accurately, make devices behave in line with user expectations, and become more intelligent, thus improving product user experience. However, in the process, there is a lot of data interaction between the edge and the cloud. Given that the shared data contain a large amount of private information, preserving information security on the shared data is an important issue that cannot be neglected. In this paper, we combine deep learning with homomorphic encryption algorithm and design a deep learning network model based on secure multiparty computing (MPC). In the whole process, we realize that the cloud only owns the encryption samples of users, and users do not own any parameters or structural information related to the model. In the experimental part, we input the encrypted Mnist and Cifar-10 datasets into the model for testing, and the results show that the classification accuracy rate of the encrypted Mnist can reach 99.21%, which is very close to the result under plaintext. The classification accuracy rate of encrypted Cifar-10 can reach 91.35%, slightly lower than the test result in plaintext and better than the existing deep learning network model that can realize data privacy protection.

Quantum Computing Based Neural Networks for Anomaly Classification in燫eal-Time Surveillance Videos

“…Deep learning using a homomorphic encryption technique has been introduced in [18] to secure data privacy whenever customers utilize the cloud-based deep learning approach. However, the complexity is high when the depth of the networks is increased.…”

P‐2.24: A Survey of Privacy Protection Research based on Virtual Reality

“…The purpose of deep learning applied to VR is to compute the data without revealing the original content. Common approaches are secure multi-party computation and homomorphic encryption [14]. However, reducing the computational burden remains a significant challenge.…”

Digital Teaching Management System Based on Deep Learning of Internet of Things

“…The connection between each two nodes represents that the input node becomes an output node but the software is built according to the needs. Therefore, the construction of online education resource management information system in colleges and universities has become the focus of research, which aims to solve the construction of an allround teaching resource management information system and realize the sharing of teaching resources, so as to change teaching methods, improve teaching level, and improve the utilization rate of educational resources, which can promote teaching management under the new situation [10]. Idrissi and others believe that usually, system testing is unit testing, that is, module testing.…”

“…Since Equation ( 9) s an inverse Lafite function (most unknowns are more than a large number of equations), in order to obtain its unique solution, it is necessary to introduce a regular term z k about the characteristic graph, and the cost function is shown in Equation (10):…”