It is hypothesised that conventional back propagation networks have limitations arising at least, from the simplicity of the artificial neuron model that is used. A more complex neuron, called a micronet, to distinguish it from the conventional neuron, is introduced. The architecture and heuristic of the micronet are described. Visual examples of the complex decision surfaces that can be produced by a single micronet in response to a range of problems are presented.
Lecture notes in computer science : engineering of intelligent systems : Proceedings of the 14th International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems (IEA/AIE 2001), Budapest, Hungary, 04-07 June 2001,
Vol. 2070 (2001), pp. 281-289