Looking for help improving our social network fake users detection algorithm

I'm working on a cryptocurrency and the idea is that we allow users to link together, then run a social network sybil detection algorithm in order to identify users that are cheating the network. Then using those IDs we distribute currency in a manner made to stabilize in value faster as well as allowing everyone to mine with equal mining power. We have the sybil detection working with very high accuracy, however there are a few improvements we would like to add.

More urgently, we have a patent violation if we were to release the code as it currently stands. However, we also have found a work around that is actually supposed to provide much higher performance. I would greatly appreciate helping making a couple relatively small tweaks to our algorithm.

I'm happy to pay for your help in the form of the cryptocurrency you would be working on.