Computing with Spikes

A frightening thought for a computer scientist is that there might be completely different ways of designing computing machinery, that we may miss by focusing on incremental improvements of current designs. In fact, we know that there exist much better design strategies, since the human brain has information processing capabilities that are in many aspects superior to our current computers. Furthermore the brain does not require expensive programming, debugging or replacement of failed parts, and it consumes just 10-20 Watts of energy. Unfortunately, most information processing strategies of the brain are still a mystery. In spite of many positive reports in the media, even the most basic questions concerning the organization of the computational units of the brain and the way in which the brain implements learning and memory, have not yet been answered. They are waiting to be unraveled by concerted efforts of scientists from many disciplines. Computer science is one of the disciplines from which substantial contributions are expected, and in fact other countries have established already hundreds of research facilities in the new hybrid discipline Computational Neuroscience, which is dedicated to the investigation of computational principles in the brain. Computer scientists are contributing to these efforts through their experience in the evaluation of real and hypothetical systems that compute, as well as experience with robots and other machines that learn, move around, and explore their