Parallel learning and regeneration of images using a structured recurrent neural network

A recurrent neural network (RNN) has been already studied for some applications and have been also demonstrated for time series. In this paper, a new structured RNN is introduced. This network is designed with some groups of neurons and it is suitable for parallel processing and for realizing chaotic data. In particular, this network is actually implemented in a parallel computer and the performance of this network is explored for image memorizing.