0

I submitted a paper a couple months ago. I wasn't super happy with the results I published and ended up rewriting major sections of the paper (mainly for clarity rather than content). I got some better results to a point where I was happy with the results. I then submitted it to another conference. I forgot I had submitted to the first conference (partly due to medical issues, partly my fault for not having a system in place considering the issues), thinking that I had not submitted since I got the better results/did the edits. Looking back, the methods aren't that different (mainly hyperparameter optimization in ML that led to better results). How do I address this double submission?

Improve this question
2
  • 1
    Are you still waiting for submission to the first conference? Whenever they get back to you, you can decline being included in the conference.
    – Parever
    Oct 3 at 22:45
  • @Parever - Is it as simple as that? No need to mention double submission or try to withdraw right now? I am waiting on the decision from the first conference, but the paper has likely gone to reviewers at both already.
    – olivarb
    Oct 3 at 23:03

1 Answer 1

Reset to default
0

You could:

  1. Withdraw the earlier paper (I'm assuming that's the one you care less about) now
  2. Withdraw if/when it is accepted
  3. Depending on how different the papers are, do nothing

You mention you have new and different results and revised the paper significantly. Is it possible that they could be considered separate works?

If not, take your pick of option 1 or 2. I would slightly prefer 1, only because it might be more respectful of the time and effort of the reviewers/conference organizers. You don't necessarily need to give a specific reason, just withdraw. I suppose if you were trying to be as ethical as possible, you might inform one or both conferences of the mixup. I'm not sure who that would benefit though.

Improve this answer
10
  • How do I find out if it passes the bar to be considered separate? I don't think hyperparameter tuning and rewriting of the paper is a high enough bar for a new paper, although I'm unsure. In other terms, if both got published, I would never expect v1 to get cited and always assume v2 was cited assuming perfectly rational behavior
    – olivarb
    Oct 4 at 6:22
  • That probably depends on the details. I would say that if all aspects of the paper have been significantly updated/changed and new results are presented (without repeating old results) the papers are different. If the 2nd paper is just a revision of the 1st, then I do not think they are different enough to leave.
    – sErISaNo
    Oct 4 at 6:34
  • 1
    Is #2 not the safest option that occurs without leaving a mark on my record? I presume the paper is mostly through review at this point, given the timelines of the conference. Why would anyone admit to double submission/how would they get in trouble for it then. I am trying to follow the ethical path but I would also not like to ruin my name without reason
    – olivarb
    Oct 4 at 6:36
  • 1
    The difference is that "version 2" papers generally mention that a version 1 exists, cite it, and explain what the difference is. I would argue that the second and third option are not ethical choices, based on the information we have.
    – lighthouse keeper
    Oct 4 at 6:37
  • 2
    @olivarb I think 1 is the safest and probably most ethical, whether or not you disclose the mistake. I would be worried (depending on the size of your field) that both papers would make their way to the same people. That is certainly the worst outcome.
    – sErISaNo
    Oct 4 at 6:41

You must log in to answer this question.

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