At the turn of the century (2006) IBM launched a project to map the human brain as a starting point to create a “cognitive computer”. It was clear from the beginning that it was an impossible dream, yet the hope was that pursuing it would increase our knowledge of the computational structures of the brain to the point that it could be possible to use the knowledge to re-create those structures in a computer (or design a computer based on those structures).
Others focussed on scaled down brains as more pursuable alternative: insects’ brain. At Opteran Technologies the decision was to look at bees’ brain. It consists of “just” 1 million neurones, yet it empowers the bee with very sophisticated autonomy (environment awareness, image recognition, flight control…).
Now Opteran is offering off-the-shelf brains to power robots (watch the clip). They call this “natural intelligence on silicon”.
Insects may not have the kind of intelligence, or awareness, that we, and other animals, have but for sure they are very successful and better than our most advanced robots (in addition they use far less power!).
What I find interesting in this approach is that it demonstrates that intelligence, at least a significant subset of it, is about structures, there is nothing immaterial about it: if you can replicate the structure (and interactions) of a bee brain you get its functions.
The Opteran off-the-shelf brain consists of 2 video cameras (providing a 360° view) and a chip. The package weights just 30g (way more than a bee! It is equivalent to 265 bees…) and uses just 3W of power, way less than control systems used in robots.
Looks like there’s a lot that we can learn from Nature!