Sensitivity and perturbation analysis of artificial feedback neural networks

Summary form only given. The authors apply the result established by A.N. Michel et al. (1992) to systems of differential inequalities. By using the differential inequalities to dominate a class of nonlinear systems, they obtain some new results on robust stability of the nonlinear systems. They show that, when applied to linear systems, these results yield robustness criteria which are considerably simpler than corresponding existing robust stability tests. However, the principal objective of the present work is to apply the above results in the study of uncertainty issues of artificial feedback neural networks, such as Hopfield neural networks. The authors attempt to provide sufficient conditions to ensure robust stability properties of equilibria of neural networks