I am a computer science undergraduate going into grad school and trying to learn more about research.

How does one proceed when publishing a research paper to refute the conclusions (or some of the conclusions) of another (someone else's) research paper? For instance, let's say that you study a machine learning research paper that makes some dubious claims, and then offers hand-wavy reasoning and a computational simulation experiment to support said claims. If you then study the subject matter further and use more formal and/or rigorous reasoning (perhaps a formal proof) to show that the conclusions (or some of the conclusions) are wrong, then how does one go about publishing a paper that refutes the other research paper? Is there anything different about this type of research paper (or, rather, should anything be done differently) compared to the more conventional, non-refutational type of research paper?

  • 2
    Once you are in graduate school, this is the sort of question to discuss with your advisor. – GEdgar Dec 29 '20 at 13:25
  • 1
    Some folks make a cottage industry out of publishing papers with titles like “A cautionary note regarding yadda, yadda ...” The line between smart and snark can be fine. – Ed V Dec 29 '20 at 14:15
  • Treat others the way you'd like them to treat you. – A rural reader Mar 29 at 0:34

Showing that something in the literature is shaky or incomplete or mistaken is part of how science advances.

If you can replace a published claim with something better rather than just showing that it's wrong, all the better.

When you find an instance of that you write it up carefully, professionally and kindly and submit it to an appropriate journal - perhaps the one that published the original work.

  • 1
    And cite the original, of course. – Buffy Dec 29 '20 at 16:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.