Publications
The work, written down and shared. Papers and preprints — open by default, because a result nobody can check is just a rumor with footnotes.
-
Validation of GPU Computation in Decentralized, Trustless Networks
Verifying computation in decentralized networks is a fundamental challenge, especially for GPU workloads: exact recomputation fails because GPUs are not perfectly deterministic, trusted execution environments require specialized hardware, and fully homomorphic encryption is prohibitively costly at scale. The paper adapts three verification methodologies — model fingerprinting, semantic similarity, and GPU profiling — into a binary reference model with ternary consensus, so a network can catch a dishonest result without redoing the work. This is the validation groundwork behind Lattice Protocol's trustless compute layer.
Read the paper →