Thank you for your detailed reply. Not here to put your ico into the ground, just looking for answers to some question marks.
Completely understand and more than encourage your due diligence, something which is often lacking in these markets!
I'm guessing you're using a neural network?
I'm curious how you will get your dataset to train your neural network.
The architecture of the distributed processing is still to be defined as such we will have a specialist write a dedicated blog post around this.
However, the preliminary plan is CTN nodes taking instruction from a master-node; the master node creates 'todo lists' of blockchain data with verified, expected results to be trained on (and to ensure we're not wasting resources by doing re-work.)
The CTN node will then process the allocated work, training on it's slice of the model in order to find relationships within the data.
The output relationships from training data will be merged across nodes on the network, from here their results are verified before being consumed by the CTN Risk Rating.