Hardware Implementation of Neural Networks

Nowadays neural networks become very useful in different fields of research such as image recognition, optimization, data analysis, classification and prediction tasks. Though, increasing complexity leads to longer calculation time and higher power consumption, what constrain using of neural networks in different types of applications. This makes a reason to develop specific hardware implemented circuits for neural networks. This paper briefly reviews present possibilities and approaches of hardware implementation for some types of neural networks.