3

I was wondering if one can submit to two conferences/journals, but one is non-archival (e.g., a workshop). My curiosity stems from the NeurIPS experiment showing that the submission process seems highly stochastic, so if the submission does pass the non-archival compliance, I wanted it to go to a workshop at least that is non-archival e.g.

Publication Accepted papers and supplementary material will be made available on the workshop website. However, these do not constitute archival publications, and no formal workshop proceedings will be made public, meaning contributors are free to publish their work in archival journals or conferences.

So, is it ethical to submit nearly the same paper twice, but one is archival, and the other is not?

Also, what would one do if it gets accepted by both?

Context

Fyi, I know two submissions to archival places are wrong, e.g., I submitted a paper to two journals, and the first published it, but I agreed to publish in the other; what can I do? but it's confusing given non-archival places like AI workshops.

Mehdi Abbassi
  • 211
  • 2
  • 7
Charlie Parker
  • 132
  • 7
  • 4
    There are almost certainly norms in your field about this. Ask someone in your field. – Alexander Woo Sep 20 '22 at 19:20
  • My did several variants of a theme, so perhaps if you change something or extend it (somewhat like what Buffy's answer says) it will work out beter. – JosephDoggie Sep 20 '22 at 20:11
  • 1
    The whole "paper count" idea as a way of assessing peoples' work leads to this (to my mind bogus) constraint of not being allowed to tell a story more than once... and, gosh, maybe it's a story that isn't over, is ongoing, etc? "Talking about one's current progress" is surely the desired thing... as opposed to having it be disallowed because you'd already talked about it? After all, big-shots get to tell their conjectures more than once, eh? :) – paul garrett Sep 20 '22 at 22:28
  • @AlexanderWoo I did and they said parallel submission to workshops are fine but of course details lie in the devils for call for papers e.g. CVPR doesn't allow it while ICLR NeurIPS workshop doesn't care. The NeurIPS workshop even explicitly says it's fine. There are already some great answers. :) – Charlie Parker Sep 27 '22 at 17:57

4 Answers4

2

The workshop clearly states that they are okay with people submitting their workshop submissions to another venue.

If you want to do that, ask the other venue for permission. Show them the quote you showed us. Then, if they are paying attention, they will grant you permission.

The main point is: Don't waste people's time reading the same submission more than is necessary.

Anonymous Physicist
  • 98,828
  • 24
  • 203
  • 351
2

This is fairly common, in my experience, at top machine learning (and adjacent) conferences. Usually, the non-archival venue is a workshop attached to some conference, so the repeat submission typically happens in one of the following ways:

  1. Early version of paper is submitted to a non-archival workshop A, and then to an archival venue B (e.g. a conference). This could also happen if the paper was first rejected at the main track of the conference which A is attached to.
  2. Paper is accepted at main track of a conference (archival), and then submitted to a workshop (non-archival) attached to the same conference, whose focus area is the subfield the paper belongs to. Some workshops may accept such papers without additional review.
  3. A previously published paper at an archival venue is later submitted to a workshop that focuses on its subfield.

The idea behind additionally submitting to a non-archival workshop is typically to allow more detailed presentations (e.g. an oral presentation as compared to "just" a poster at a larger conference), more detailed discussions, more visibility for the paper, and to an audience that primarily is from the same subfield (as compared to a broader audience at large conferences).

I am not sure if submitting to an archival and non-archival venue simultaneously is common, but I do not see any ethical issues, beyond two primary considerations:

  1. As other answers say, you should check the policies of the archival (and non-archival) venue. For example, NeurIPS allows submitting papers that have previously appeared in non-archival venues:

Work that has appeared in non-archival workshops, such as workshops at NeurIPS/ICML, may be submitted.

CVPR, on the other hand, does not:

A publication, for the purposes of this policy, is defined to be a written work longer than four pages (excluding references) that was submitted for review by peers for either acceptance or rejection, and, after review, was accepted. In particular, this definition of publication does not depend upon whether such an accepted written work appears in a formal proceedings or whether the organizers declare that such work “counts as a publication”.

Additionally, some venues may allow such submissions in general, but may restrict them during their review period.

  1. If the submission to the non-archival venue is after the archival paper has been published (with copyright transfer), you might need to check the policies of the publisher regarding this.

If it gets accepted by both, you will get the chance to present at both. Formal proceedings and copyright transfer would only take place for the archival venue, and that version would typically be considered "canonical" when listing on publication lists and for citing the paper. In particular, assuming the two are nearly identical, you do not get to count them separately in your list of publications.

The question of submitting to arXiv is a orthogonal to this discussion, as arXiv is an archival "venue", more than most non-archival workshops, but by not being peer-reviewed, does not have the ethical concern of wasting reviewer time or questions regarding copyright transfer.

GoodDeeds
  • 4,036
  • 6
  • 21
  • 40
  • I think it's worth emphasizing that it seems CVPR doesn't allow workshop publications but it does allow arxiv "publications" (since it's not peer reviewed) -- since both are none archival. Thanks! – Charlie Parker Sep 26 '22 at 18:45
  • 1
    @CharlieParker I would actually say arXiv is archival, and particularly more than most non-archival workshops, since arXiv publications have a unique ID and are expected to be persistent, while non-archival workshops (at the most) upload the paper on their websites, with no guarantees that it will still be available in the future. This is why non-archival workshop papers are often cited using their arXiv identifiers. However, arXiv submissions for the most part are not considered when talking about dual submissions in CS, and the ethical argument is primarily about not wasting reviewer time. – GoodDeeds Sep 27 '22 at 04:50
1

This may be a distinction without a meaning. The workshop is telling you that they make no claim, copyright or otherwise, on the material so won't bind your submissions elsewhere. However, it is a form of "publishing".

And if you submit to a journal it is their rules that will apply to whether it will be published there or not. Some journals won't publish things "published" elsewhere, though that is changing and has changed in some fields. If you submit to such a journal after "publishing" the same work elsewhere, they will be unhappy.

There are no laws (that I'm aware of) for such things, but the journals and conferences make their own rules.

It might be safer and more appropriate to submit the paper to a journal snd submit something different, but related to the conference workshop.

Workshops, being more informal, probably accept partial work and work in progress. The point is sharing ideas, not presenting finished work. That is my experience anyway. You can certainly talk about a paper at a workshop with no consequences.

Said another way, the conference gives you permission to publish elsewhere, but a journal may not. It isn't necessarily symmetric. Caution is advised.

You can, of course, check with any journal you propose submitting to. Contact the editor and/or read the submission guidelines. If they permit preprint publication (arXiv...) then it is likely they will permit what you would like to do. (likely, not certain)

Buffy
  • 363,966
  • 84
  • 956
  • 1,406
0

For this specific NeurIPS workshop:

  1. Can a paper be submitted to the workshop that is currently under review or will be under review at a conference during the review phase?

From our side, it is perfectly fine to submit a condensed version of a parallel conference submission if it is also fine for the conference in question. Our workshop does not have archival proceedings, and therefore parallel submissions of extended versions to other conferences are acceptable.

Charlie Parker
  • 132
  • 7