An Aston University led consortium has won a funding of 3.5 M€ from the European Commission to explore the possibility of using natural neurones (and neurones networks) in conjunction with silicon chips to improve computation resulting in artificial intelligence.
Now, this goal is not just ambitious, it is in the realm of science fiction, but that is just how the project has been explained by journalists (and for journalists). The reality is much more limited, and yet it is very ambitious (there would be not need to bloat reality…).
The 3 year project leverages on multidisciplinary expertise from biology and computer science to study neurones and their possible interactions with silicon. Stem cells that are specialised to become neurones are layered on silicon embedded in a nutrient solution. Human neurones (like any other animal neurones) are relatively slow in processing (their clock speed is measured in ms, the clock speed of a silicon chip is measured in ns, one million times faster) but they create an amazing parallel computing processes and storage (it is difficult, if not impossible to separate storage from processing in neurones) and this parallelism is both a consequence of the billions of neurones in our brain as well as of the thousands of synapses in each neurone giving rise to trillions of connection/computation gates. A point in favour of the brain computation is the very low energy usage, orders of magnitude less than the silicon in a chip.
The goal of the Neu-ChiP (this is the name) project is to create a hybrid, symbiotic structure that can leverage the strength of silicon and carbon based computation, of chip and neurones.
The first step is to understand the plasticity of neurones networks (the way they interact as well as the way they form and change connections as result of external stimuli) by firing changing patterns of light on the neurones (optogenetics) and using sophisticated 3D modelling software to study the changes.
Most likely the project will advance our understanding of the computation underpinning of neurones and that will be a step forward to create neuromorphic chips that mimic our brain. The symbioses of human brain with computer brain for several more years will continue to be mediated by our senses and computer interfaces. Do not under-estimate this symbioses: we are already in a sort of symbioses with our smartphone that in a way has become an extension of our memory.