It is important for a research paper to include raw data and code for scientific replicability, verifiability, and falsifiability.

However, as noted by nbro, roughly 5% percent of papers have code that precisely reproduce the experimental results. About 20% of papers might have implementations coded by other researchers, but they usually differ from the original code and implementation of the author.

Is it possible to find AI research papers where the author provide source codes aimed to reproduce the experiments described in the paper and produce exactly the same plots?

  • $\begingroup$ Wait, note that I didn't say for sure that only 5% of papers include source code. What I said or at least meant is that, from my experience, I think that only roughly 5% of the papers have code that you can use to really reproduce the experiments in the paper (but this estimate could be wrong - it's just a very very rough estimate). Maybe I should have been more precise in that comment. Please, edit your post to fix that. $\endgroup$
    – nbro
    Jul 10, 2022 at 11:28
  • $\begingroup$ @nbro Does it sound better now? $\endgroup$
    – High GPA
    Jul 10, 2022 at 12:07


You must log in to answer this question.

Browse other questions tagged .