This paper has been accepted, presented in the top conference of computer science, and released in the proceeding since last month. However, recently, other readers find a fatal and ridiculous error in the proof. For this fatal error, I think I cannot amend it. Even if it can be amended, some important claims and nearly all proof steps will become total different.

Something to stress:

1, My claims and other proofs are motivated by the wrong proof rather than the experiments.

2, I then used experiments to demonstrate the correctness of the claims.

How can I handle it?

  • 1
    Just explaining the edit: When you withdraw a paper after it's published in the proceedings it's called "retraction" (before publication, it's withdrawal).
    – ff524
    Jan 17, 2017 at 3:23
  • 1
    Have you examined the possibility of proving the other results without using the result with the flawed proof? As I understand from comments, the claims do have experimental support, even without any of the proofs. Jan 17, 2017 at 8:48
  • Why have you edited your question to remove practically all of the content?
    – Pont
    Jan 20, 2017 at 12:55
  • 1
    I have restored the content of the question. Questions and answers should be available for future visitors.
    – mhwombat
    Jan 20, 2017 at 13:08
  • 2
    @olivia We all do stupid stuff from time to time - that's ok. What matters is how you handle it.
    – Dirk
    Jan 20, 2017 at 14:02

1 Answer 1


For a journal paper you could either retract it, which means to "unpublish" it, or you could add an erratum, i.e. a short note where you explain what parts of the paper are wrong (and how to fix them, in case you know it) and which parts are still correct.

If I understood your case correctly, it seems like an erratum stating that the mathematical proof is incorrect, while the experiments still support the claim, would be appropriate.

You must log in to answer this question.

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