The Defacto Standard, page-20

That would be a good symbiosis with AKIDA.

Machine Learning has its place - think of it as a collective memory - taking in data from all over the world and building AI algorithms off it.

AKIDA can use that as a baseline, then use its on chip learning to tailor the algorithm to your use case.

Example: Voice recognition can be a bit ropey if you have a Kiwi accent (even one that is as Australianised as mine) That's because the voice recognition algos work predominantly on US accented datasets.

So Akida gets the ML voice recognition algo as a "cheat sheet" - then adjusts it to account for my trans Tasman, slightly British accent.

Or maybe you have a darker complexion/skin tone - a lot of the facial recognition software struggles with that - because the data sets are based on a whole bunch of white people. Same same, AKIDA can take the baseline algo and tweak it a lot faster than sending that data back and forth to a ML server.