Exploring ZK Proofs for ML Computation in Blockchain

Just launched a blockchain that needs zero-knowledge proofs to confirm block validity based on ML computations. Thought I'd share here for some tech feedback since it relates to your interests. So here's how it works. Each block includes a hidden Markov matrix, M, and the block hash represents an observation sequence, O. Miners are tasked with figuring out the most likely state sequence, s*, using the Viterbi algorithm. Then they need to generate a Schnorr sigma proof to show that s* is accurate based on M and O. For the proof structure, I'm using the BN128 curve: - There's a generator G and a blinding generator H calculated as scalar_mult(G, H_seed). - For each state transition, I'm doing a Pedersen commitment C = scalar_m. Would love to hear your thoughts on this!

Could you share more about the implementation? Looks complex.

What if everyone just takes the most probable state? That could screw things up.

I see your point, but if it’s done right, it should mitigate that risk.

So you built this yourself? Impressive!

Yeah, I'm all about experimenting with new tech.