Great work @crokkon! A very clear case.
I hope to see more of this kind of contribution in utopian antiabuse. With all the data transparent and available from the blockchain we should be able to make good use of pattern recognition tools to uncover persistent abuse.
Totally agree.
I was wondering if this is something ML (Tensor Flow and the like) can be applied?
I think a big data approach. Throw all the blockchain data into a pool. See what patterns past abuse cases throw up. Use these to find and predict new abuse at early stages.
Thanks @miniature-tiger! In this case, it was indeed very clear. However, not all are that clear. Having the blockchain data helps a lot, though. Pattern recognition is an interesting approach, and also ML as mentioned in @lovenfreedom's comment. I tried once to analyze only a few days in the context of "who votes mostly together with whom" in a brute-force approach - This gave a few interesting results, but the processing time requirements would not have scaled to larger time ranges. Applying ML/AI techniques could indeed help there to build models...