A frame based implementation architecture for neural networks

Neural networks have the potential to provide very cost effective pattern recognition machines provided that suitable hardware implementations can be found. The applicability of common neural network structures, such as the multi-layer feed-forward network, to different pattern recognition problems means that any particular implementation scheme will be widely applicable. This generality makes it work seeking implementation schemes which are able to provide the neural network designer with a flexible building block and the system designer with an efficient component level structure. This paper describes an implementation architecture that has been designed by the authors to fulfil these requirements, and is called the data frame architecture (DFA)