New Generation Encryption Chip Based on the GAN Chaos Model and the Geffe Architecture

In this paper, we proposed a new generation encryption chip based on the generative adversarial network (GAN) Chaos model and the Geffe architecture. Compared with software encryption, hardware encryption is not easy to be attacked, data and programs are mixed, and hardware cracking analysis is difficult. It can effectively avoid attacks such as reverse engineering and the calculation speed is faster. In order to perform more unpredictable hardware encryption, this study designs the GAN algorithm to generate encryption circuit, establishes the convolutional neural network (CNN) model through training, and combines the random number generator circuit of Geffe architecture to realize the system with digital integrated circuit. To reduce and improve the calculation speed, the iterative calculation method is used to design, which can reduce the circuit area, optimize memory consumption, improve the calculation speed of CNN, and achieve a faster AI encryption inference chip. The experimental results prove that our proposed method reduces the computing time by 78.4% and realizes a high-performance encryption chip.