A VLSI Implementation of Novel Neural Computing Unit

Since the application of artificial intelligence in speech recognition, image processing and other directions has become more and more widespread, the requirements of computer computing efficiency and power consumption performance have also been continuously increased. The main solution is combining traditional algorithms and artificial intelligence to solve this problem. However, this method cannot fundamentally solve the problem of insufficient computer performance. At the same time, due to the fact that the widely used neural units do not have enough biological neurological properties, artificial neurons do not have the ability of autonomous learning. In view of these issues and the information transfer process of biological neurons, a new hardware architecture for the basic neural unit in the neural network was proposed in this paper.