Bifurcating pulsed neural networks, chaotic neural networks and parametric recursions: conciliating different frameworks in neuro-like computing

This paper addresses the concept of parametric recursions in the context of neuro-like architectures and relates it to different classes of neural networks, including pulsed models and chaotic neural networks. It also addresses the contrast between the dynamical states exhibited by networks of parametric recursive elements and the fixed point attractors used in traditional neural networks, as well as presents results from neural architectures employing parametric recursions and coding of information through dynamical attractors.