The SpiNNaker Project

----> a universal Spiking Neural Network architecture <----

The SpiNNaker project has been awarded £ 1Million by the Engineering and Physical Sciences Research Council with the aim to build a computer which mimics how nerve cells in the brain interact in a bid to engineer more ‘fault tolerant’ electronics. The work will be carried out in collaboration with the School of Electronics and Computer Science at the University of Southampton, using technology supplied by industrial partners ARM Limited. and Silistix Limited.

The classical computational paradigm performs impressive feats of calculation but fails in some of the simplest tasks that we humans undertake with ease and from a very early age. Biological neural networks are proof that there are alternative computational architectures that can outperform our fastest systems in tasks such as face recognition, speech processing, and the use of natural language. Brains are complex highly-parallel systems that employ imperfect and slow (though exceedingly power-efficient) components in asynchronous dynamical configurations to carry out sophisticated information processing functions. Note the word asynchronous in the previous sentence! Many aspects of brain function are little-understood, but we hope that our deep understanding of the engineering of complex asynchronous systems may be of use in the Grand Challenge of understanding the architecture of brain and mind.

The real-time modelling of large systems of spiking neurons is computationally very demanding in terms of processing power, synaptic weight memory requirements and communication throughput. It is proposed to build a high performance computer for this purpose with a multicast communications infrastructure inspired by neurobiology. The core component will be a chip multiprocessor incorporating some tens of small embedded processors, interconnected by an NoC which carries spike events between processors on the same or different chips. The design will emphasise modelling flexibility, power efficiency, and fault tolerance, and is intended to yield a general-purpose platform for the real-time simulation of large-scale spiking neural systems.