Big data with its associated acquisition, analysis, and decision-making is driving the growing demand for computing power that classic processor architectures can no longer provide. That is why Intel has been working for the last six years on machines that function like the human brain – which may be " poor in terms of calculating, but good at everything else ". One result is the AI chip Loihi , which initially is available only for universities and research institutions. This extremely energy-efficient Intel development will gradually become more intelligent by itself and no longer has to be trained as used to be the case. To do so, it takes a novel approach to calculation: its "neuromorphic computing" is based on our current understanding of how the human brain functions. The neural networks relay information as pulses or spikes, modulate the synaptic strengths or weight of the connections based on the timing of these spikes, and store these changes locally at the connections.
The Human Brain Project (HBP) has been running in Europe since 2013. Scheduled to continue for 10 years, it involves 112 partners in 24 countries. It, too, involves neuroscience-inspired supercomputers (as well as neurobotics ). The University of Manchester’s SpiNNaker Project and above all the completed research project BrainScaleS (brain-inspired multiscale computation in neuromorphic hybrid systems) by Prof. Karlheinz Meier at the University of Heidelberg, which also works with digital neurons and synapses, have to be mentioned in this context. Brain-inspired computing is now being taught in Baden-Wuerttemberg. For Meier’s team in Heidelberg a new research building is being built: the European Institute for Neuromorphic Computing (EINC).