For example, for a novel artificial intelligence algorithm, one may have some description of the experimental data and results, in addition to the description of the artificial intelligence algorithm itself. By experimental data and results, I mean a systematic analysis of the results (e.g., reporting some performance metrics on an entire test set) and not just some cherry-picked examples. Is there any point in including the description of the experimental data and results in the patent application?

I see some patents adding the description of the experimental data and results {1}, and some that don't {2}.


  • 1
    Congratulations on your patent issuing.
    – Eric S
    Jan 8 at 15:50
  • @EricS thanks! Most of the credit goes toward my collaborators. Also we still have a bit more work to do to reach your number :-) Jan 8 at 20:10

2 Answers 2


The U.S. a requirement to describe how to make and use the invention. There is also a “written description” requirement to show that you “possess” the invention. That means you understand that it works without any further invention. It may be that one person filing an application is more concern about meeting that hurdle than another applicant or one invention is less believable on its face, requiring more demonstration of results.


There are several aspects to the requrements that a patent specification must fulfil. Three in particular are specified in 35 USC 112(a):

"The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention."

Out of the three requirements, for written description, enabling disclosure and 'best mode', the existing previous answer has already covered the first of them. It might not be totally clear to every inventor or draftsperson how to assure in any given case that the others, the 'enablement' and 'best mode' requirements, are also satisfied. Giving experimental data and results as well as workable algorithms can be a way to make sure that these requirements too have been fulfilled.

There are sometimes arguments against, as well as for, a description that could be unnecessarily full, but often, caution towards completeness is the better policy.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .