The remarkable computational power of the brain has motivated researchers to investigate computational systems which draw inspiration from the nervous system whose exact workings are still very much a mystery. Several neural models and encoding schemes were proposed and a number of learning rules were formulated. However, their applicability has been hampered by the lack of efficient training mechanisms. This work treats the idea of using bio-inspired neuron and synapse models in developing computational systems that can be applied to solve real world problems. A review of traditional ANNs and the more biologically plausible ANNs, i.e. spiking neural networks or SNNs, is presented. Different aspects and variants of evolutionary algorithms are addressed. Alternative training approaches/architectures based on evolutionary strategies, dynamic synapses and neural oscillators are developed and their performance and scalability are evaluated. This work should help shed some light on the growing field of SNNs and their design and application challenges. This book should interest readers, students and researchers, in neural networks, intelligent systems, computer science and engineering.
