Building Smart Machines by Utilizing Spiking Neural Networks; Current Perspectives

In this paper we survey the current state of the art in spiking neural networks research and outline our approach to building smart machines. A thorough understanding of the history, open questions, and limitations of these networks can help the research community to gain a better grip on this new technology and to bridge the missing gaps. It is necessary to look at various aspects of spiking neural networks, such as the different modeling approaches, encoding schemes, simulators and learning techniques in order to efficiently make use of these networks. One paramount characteristic of spiking neural networks is the precise timing of inputs and outputs. As a dynamic system itself, it naturally lends itself to solving problems in the continuous domain such as time series analysis. This will be the focal point of our efforts to develop a smart machine utilizing spiking neural networks