The Science of Certainty
K-Veritas seals each result to the exact code and moment that produced it.
Change either afterwards and verification fails. Even an attempt to fake it gets sealed.
It proves what was run, not that the experiment is right. That judgment stays with you.
Your data is never accessed. Your code is never re-run.
Official implementation of Computer Science Conferences Should Require Nonrepudiable Experimental Results (Keita and Homan, NeurIPS 2026 Position Paper Track).
“What is written in the paper should be what was truly gotten from the experiments.”
Reports verified
Trusted by
