2

In the ML word, it's a common practice to upload a paper to Arxiv when it is still under review. But is it okay to share it with other people?

Let's say someone writes a paper for congress. And while the paper is under review, he decides to share it with other researchers of the area.

Is it ethically fair to do that?

9
  • 16
    They put it on Arxiv - that is sharing it already.
    – Jon Custer
    Jul 30 at 14:23
  • 9
    It isn’t like Arxiv is secret. Others in your field are perfectly capable of searching for it as well. The whole point of Arxiv is to share preprints quickly and easily.
    – Jon Custer
    Jul 30 at 14:36
  • 1
    I see your point! Many thanks! Jul 30 at 14:39
  • 3
    If a paper isn't ready to be seen by your enemies, it isn't ready to submit to the journal yet! If an enemy sees your paper on Arxiv (or elsewhere) and send you their criticism, they are really your friend. Much better that than leave the job to reviewer 3! ;o) Jul 30 at 14:41
  • 2
    @CarlosMougan The point of double blind reviews is that the reviewers should not know who the authors are. Some places consider that a preprint on arxiv already breaks this anonymity. Because allowing early preprints is desirable, others make some compromise, where it is ok to have preprints and the reviewers are encouraged not to try too hard to identify the authors. But that becomes hard if the author contacts people or gives a talk or something that draws extra attention, so some places try to limit it... So read the specific guidelines when you submit. Jul 31 at 7:57
1

I won't comment on the ethics, but I want to point out that by submitting your paper to a conference, you agreed to some terms, and it would be a good idea to check those terms. Those terms may have constraints on what else you can do with the paper, either for copyright reasons, or more interestingly to ensure some anonymity in double-blind reviewing.

Having a copy of your paper on arXiv makes it quite easy for a reviewer to uncover your identity. Some conferences with double-blind reviewing thus forbid anything that can help identify you (including uploading to arXiv) until the end of the review period (at least). Some allow anything and trust reviewers not to go looking for the information. And others have some intermediate policy, for instance allowing a preprint on arXiv only if it was uploaded long enough in advance (so it doesn't show up on news feeds during the review period?), or forbidding to give a public seminar about the paper (this could also be to ensure that the conference gets some kind of exclusivity, is the first place where the paper is presented, but I have seen it argued as a way to preserve anonymity). If you personally send your paper to some people and one of them happens to be a reviewer...

6

It's common and ethically just fine to share work with specific others at any stage of research, even prior to the advent of preprint archives/in fields where preprints remain uncommon. It's your work. You can share it to get feedback, to inform others of work you've already done so that they don't waste their efforts, etc.

I would not recommend sharing work unsolicited broadly to people you don't know. That's less about ethics and more about politeness around unsolicited communication. There may be cases where you want a specific individual, preferably someone you have already met, to give feedback on your work; then it's fine to share with them (you may want to ask first something like: "I am working on XYZ, would you be willing to have a look at it and offer your thoughts?"). If you want to share work broadly, however, the way to do it is with a preprint. It's fine to point people to that preprint through other means, like on a social media account, but don't spam them; that's not unethical, but it is rude.

Exceptions: patentable work is legally complicated. Don't share such work without understanding the legal implications. Additionally, if you are under any additional contract terms for the work you're doing make sure you are aware of those terms, like if you are being funded by a private company.

2
  • 2
    Well, don’t put patentable work on Arxiv. Never had to have that discussion with the patent lawyers, but I’m pretty sure I know their take on it…
    – Jon Custer
    Jul 30 at 14:38
  • 1
    @JonCuster "Don't put it on Arxiv" I think is pretty obvious. I'm less aware of what the implications are of sharing all or parts of the work with other outside individuals, and I assume it depends a lot on jurisdiction.
    – Bryan Krause
    Jul 30 at 14:45
4

Until you give up copyright you can share with anyone you like. Even after, informal sharing is unlikely to raise the hackles of the new copyright holder. They will often give you a license for such things, in fact.

But if it is publicly available on arXiv, just point them to it there.

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.