Convergence for a class of delayed recurrent neural networks without \(M\)-matrix condition (Q732116)

Recurrent neural networks (RNNs) have an important role in signal and image processing, pattern recognition and optimization. Many researchers focus their attention on the analysis of RNNs's properties and obtain many results. However, if you analyse these results, you can find that these results are based on nonsingular M-matrix theory. In this paper, the authors study the convergence for a class of delayed recurrent neural networks without the M-matrix condition by using different methods, and obtain more general results than before. Whether their results in this paper have applications, such as signal and image processing, pattern recognition and optimization, I could not find.