Adaptive modelling for cognitive structures

Summary form only given, as follows. The authors discuss and demonstrate adaptive modelling neural networks for use in cognitive, or thinking, systems. A simple cognitive model is presented which relies on the ability of the network to model its environment. A neural network structure is developed and used to model the signal production process behind several one-dimensional time functions, including deterministic and nondeterministic signals. The results demonstrate the ability of very simple neural structures to approximate, predict, and model the behavior of many temporal processes. This dynamic modeling process is shown to be a critical feature for autonomous systems operating in unstructured and nonstationary environments.<>