3

I am an undergrad who worked at a computer vision related lab. The lab I used to work at raised some red flags by saying that even if I contributed, my name could not be on the paper. I also transferred schools, so I left. The problem is, I have this potentially great research idea that I'd like to test. I still have the data (a one hour video) and was either a. going to ask if I could use it for solo research, of course crediting them b. try to find a new professor at my school to supervise me.

I decided I want to find another professor to supervise me, cause I have never written a paper before and this is kind of a big project. I have been reading papers and research this for 5 months now, and I have never seen my idea done before. It basically takes something from a different area of the field and applies it to the research area I am interested in. I have seen it tried before, but not in the exact way I plan on doing it. My concern is, what if I email professors about this and they steal me idea? Sure, if your ethical you won't, but could that happen? I want to email this professor with my idea, and ask if has any data / ways I can explore this at their lab. What if he says no and runs with my idea? either way, I am hoping he can offer me some perspective, but even if he says no he can be wrong - and isn't that the point of research to explore things that may not work? Another reason I want to work with him is because all though there is another professor who would be able to help, he publishes a lot in IEEE and Civil engineering journals. I am trying to get noticed by the AI/ML and deep learning community, and that is where this other professor mostly publishes in. Would it be bad to ask him if he knows of any professors I can work with, if my idea doesn't suit his lab?

6
  • Is the new professor at your new school?
    – Dawn
    Sep 2 '20 at 18:02
  • 1
    A suggestion here is to write it up as a white paper. It does not have to be written as an academic paper; just outline the problem and algorithm, and some results. Post it to ArXiv with your name of course. After that, point a targeted professor at the white paper, and ask him/her whether he/she is interested in turning the work into a proper academic paper. Sep 2 '20 at 19:13
  • 2
    Be aware that ideas are easy to find and hard to exploit. Reading your post it is hard to know where you are between "hey the technique of that other paper looks cool" and "I know what kind of results I would get but I need the paperwork before pushing the button".
    – UJM
    Sep 3 '20 at 12:49
  • no the professor at my school isn't new, I meant find a new professor to work with at my current school
    – Emilio
    Sep 6 '20 at 18:42
  • 2
    Please refer to this for deleting your question. Since there is no upvoted answer, simply deleting it may be possible (I am not sure), and you could try that. But please do not keep the question and deface it.
    – GoodDeeds
    Sep 16 '20 at 21:07
2

Unless by mischance you are contacting people who are not actually respectable (on the world stage), the sorts of issues that worry you are non-issues.

The more genuine issue, assuming the above, is that more-expert people will not immediately have the same enthusiasm for your ideas as you do. For many possible reasons, including that they and their group have already thought about that, maybe many times, and/or they are aware of other peoples' thinking in such terms decades earlier, etc.

So don't be offended, and don't be suspicious, if/when people do not hail your ideas as revolutionary, etc.

In real life (in academe, in my 40+ year experience), although, yes, sometimes unscrupulous advisors and others do indeed steal their advisees' ideas/work, ... this is (mercifully) quite rare. In particular, again, unless you are unwittingly involved with nasty losers, this worry should by far not be your first consideration. Indeed, thinking in such terms would corrupt your thinking in many ways.

And, in any case, as Michelle Obama recently said, "when they go low, you go high".

Obviously morally ideal, ... and, yes, not always easily feasible in reality... but thinking in those terms is a much better base point than thinking in terms of endless dishonesty.

2
  • "Michelle Obama recently said, 'when they go low, you go high'." - A very relevant statement for many posts on SE Academia. Sep 16 '20 at 22:51
  • "sometimes unscrupulous advisors and others do indeed steal their advisees' ideas/work" - we do not know how rare that is. SE attests that it does happen to people, and that some people seem to attract trouble somehow. Dishonesty happens, and no matter how rare, it always is upsetting, like getting mugged. You will always remember the one time. A motto that sometimes works is: "develop more good ideas than people can steal from you". Another one is: "Make your idea initially sound dull enough that, by the time someone discovers that they should have stolen it, you've already got the credit." Sep 16 '20 at 23:19
0

Let me start this by saying that the scenario you describe is not very likely. If you have an idea, done the research, and have some preliminary models, then any semi-reputable professor (at the very least, in the United States) will not steal your idea. This is due to several reasons:

  1. There are real repercussions to their reputation if you complain (say, by reaching out to student representatives or your institute's ombudsperson).
  2. You are willing to do all the work and have made a lot of headway, all they need to do is simply supervise you and they get a free publication. Why would they steal it and then struggle to go at it alone/wait til their own students figure it out? ML, and vision in particular, is an extremely fast-moving field; for all they know, if they reject you, you'll find someone else who'd be willing to work with you and they'd be wasting their efforts.

That said, if you are genuinely concerned, the best way to avoid having someone steal your idea is to write it down yourself and publish a short summary. Now, I do not mean that you must go through peer review; rather, write down your general idea into a human readable format, and then post the draft to ArXiv, where it'd be timestamped and immediately recognized as yours.

Good luck!

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.