Embedded tutorial - Can silicon machines match the efficiency of the human brain?

Summary form only given. Some recent developments from the field of nanotechnology that lay the stepping stones to deliver on this challenge will be discussed. It has been demonstrated that nanoscale devices made of chalcogenide alloys, Ag2S, TiO2 etc, could be programmed with picojoule level energy consumption to mimic the plasticity properties exhibited by biological synapses. Small prototype circuits based on spike timing plasticity have also been developed, capable of sequence learning, pattern recognition etc from noisy spatio-temporal data. The technology requirements and targets to be met by novel devices in order to fabricate large, energy-efficient, bio-mimetic computational systems will also be discussed