Convolutional Neural Network (CNN) Accelerator Chip Design

Abstract

With the development of artificial intelligence, artificial neural network has been applied in many industry fields. The convolutional neural network (CNN) which is one of the most important algorithms in deep learning plays an important role in computer vision and natural language processing. With machine learning becomes more complex, the amount of data and the amount of computation in CNN increase dramatically. A large amount of data multiplexing consumes a lot of data handling time for the traditional CPU (Von Neumann Architecture and Harvard Architecture). The data processing speed affects the CPU performance. Increasing computation speed and reducing data multiplexing have become the primary goal of neural network accelerators.