Patent: Blockchain for Open Science

US 10,659,239 B2: Blockchain for Open Science Research

This was the first patent I wrote at IBM Research. Despite a slightly hyperbolic title, it’s not patenting the idea that you can use blockchain for science. I’m not even sure what that would mean.

This is actually a patent that describes a way to write the outcome of a statistics program (like SPSS or R) to a blockchain. This is a super important step in keeping any kind of statistical analysis.

Most statistical analyses tell you how likely a certain scientific outcome is. If it’s unlikely, we usually conclude that the reason it came out that weird way is because of an experimental intervention. But, importantly, unlikely outcomes are only unlikely if run your analysis exactly once. If you just sit there rolling dice until you get lucky, it’s always possible to get a cool, unlikely seeming result. Importantly, most analyses assume that you don’t do this.

But this is impossible to enforce, we can’t sit there watching every grad student run multiple analyses and correcting their results for the number of times they run it. That would take some kind of vast network of interconnected computers and an immutable ledger—oh wait, yes we totally can. We just need to connect their statistics program to blockchain. And that’s what this patent was for!