33

I graduated with a B.S. degree and I have been working in industry for a few years as a software engineer. I am very interested in reinforcement learning, a sub-field of machine learning, and I am interested in performing research and publishing a paper on that topic.

The problem is, in my city there are no professors working in this field (or related fields) who could help me. I contacted other professors in different cities and they rejected my help request for a variety of reasons (such as I am not their student, they don't know me, they are busy, etc.).

Still, I want to perform research and publish a paper. I can't attend school as a full time student and my collaboration requests were not successful. Given all that, how can I publish by myself?

  1. How to choose a research topic?
  2. Should I read all the papers on this field?
  3. How to decide which problem to work on or which is not obsolete? I don't want to work on something which no one is interested in.
  4. Should I read and study all the prerequisite material, or fill in the gaps during research efforts?
  • 6
    I think the first question you'll have to answer is, why do you want to publish? What goal are you hoping to achieve? Once you answer that, we can probably give you more directed advice. – eykanal Sep 11 '12 at 18:30
  • 8
  • 3
    The title and the question are unrelated, and I think you confuse conducting research and publishing stuff. – Sylvain Peyronnet Sep 11 '12 at 21:09
  • 1
    I think this question is sufficiently different from those cited as duplicate: he is asking about his options for research/collaborations or independent research, very much distinct from "can I publish a paper without affliation". – Legendre Sep 12 '12 at 12:20
  • 2
    I agree with @Legendre; the linked questions address the question of "can it be done?", this question asks "how can it be done?". – eykanal Sep 12 '12 at 15:46
22

Is it possible to publish a paper without a Professor?

Sure. Nobody gives a rat's patootie about the academic ranks of paper authors (at least in my field). I published sole-author papers as a graduate student. I have colleagues who published as undergraduates, and others who published with no university (or research lab, or corporate) affiliation at all.

How to choose a research topic?

Choose something you're good at, that you're passionate about, and that other people will care about. If you're not good at it, you'll never get anywhere. If you're not passionate about it, you won't put in enough effort to succeed. If nobody cares, then, well, nobody cares.

Should I read all the papers on this field?

No. Reading all the papers in any field is simply not possible. But you should read a lot. A few hundred papers should get you started (ha ha only serious).

How to decide which problem to work on or which is not obsolete? I don't want to work on something which no one is interested in.

Read lots of papers; talk to lots of people; go to seminars/conferences and listen to talks. Or decide that the topics that you care about are so fascinating and your results are so compelling that you'll make other people interested (but be prepared for disappointment). Or—my personal favorite—just make up something cool out of thin air.

Should I read and study all the prerequisite material, or fill in the gaps during research efforts?

Yes. Both. (Except not all the prerequisite material; that's impossible.)

  • 3
    I have heard people say that for those remaining in academice after their PhD, professor-less publications are a good thing to have as they indicate you stepped out of your advisor's shadow. – Raphael Sep 13 '12 at 7:46
  • 1
    I would like to comment that "ask > read". It's easy to miss an important paper (or idea) when just searching for papers plus it's hard to decide if a problem is of the interest of others (it might be technically correct and novel, but considered as unimporant). – Piotr Migdal Sep 13 '12 at 13:59
  • This is very true. I know of an IBM employee who publishes about star formation in astronomy journals; he's not associated with a university. – Geremia Apr 19 '14 at 0:48
  • 1
    @Geremia But he still has an affiliation. What about freelancers, or unemployed people? – qed Jan 15 '15 at 21:13
10

I can't comment on the later issues—they are too general for the scope of a single question.

To address the first issue, yes, it is possible to publish without a professor. People at corporations and small companies publish all the time without academic collaborators (and without PhD level staffers). The challenge is having a topic that is worth publishing, and finding an appropriate venue to publish it in.

8

Is it possible to publish a paper without a Professor?

Yes, sure, why not? The point is whether the paper (your work) is worthy enough to be published or not. It's not the point if the paper has a professor or not; it's about being professional and not about a professor.

How to choose a research topic?

First, it depends on your interest. Then it depends on if it's useful, uses new method, new knowledge, etc.

Should I read all the papers on this field?

Not all but only related to your topics. Because you have to "re-search" if others have done it already. You need their results and methods to compare with your work for reference.

How to decide which problem to work on or which is not obsolete? I don't want to work on something which no one is interested in.

I give up for this Q. Actually,it's not hard to answer but not easy too. Because there's no right or wrong answer, only depends on your like. People select scientific papers by their own standard and so do I. No one knows exactly which standard criteria should be decided, but at least you may know the trend of interests, so please "re-search"!

Should I read and study all the prerequisite material, or fill in the gaps during research efforts?

Oops! I have to do it a lot.

PS: This may not the most correct answer because I'm not a professor but at least I hope it helps.

3

Based on your comment above, that you "want to work on a problem and solve it", I strongly recommend that you change your goal from "publishing in a journal" to simply "solving a problem". By limiting yourself to research that would be publishable in a journal, you place yourself under the following very harsh constraints:

  1. You must become familiar with the existing academic literature such that you can properly cite other academic sources when discussing prior findings.

    If you don't take the journal route, you can become familiar with existing techniques through books, tutorials, wikipedia, and blog posts, and work from there.

  2. When publishing to a journal, your work will have a higher likelihood of being published if it relates to the topic du jour. Certain concepts go in and out of style, and researching a less popular topic can have an impact on when you can publish. On a related note, there's a lot of time between submission and actual publication; a number of months to almost a year is common.

    Alternatlive, if you go your own route, then you won't have that delay.

  3. You must be willing to work in areas which are of interest to whatever journal in which you wish to publish. This requires knowledge of the different journals and what they typically publish, which may be difficult for you to find without academic contacts.

    If you don't take the journal route, you can simply publish your findings in a blog post or other open setting (there may be places specifically intended for this sort of thing, I'm not sure).

  4. You will have to pay non-trivial fees for publication.

  • I want my contributions to be notable. And the best way of this is publishing in journals or conferences. I don't see a way my work is recognized or respected via a blog post. At the end if I achieve something I want to get proper credit for it. My main target is conferences and their short paper sessions. Mainly I will target ICML and confs. one tier lower than that. – teenage ninja turtle Sep 12 '12 at 17:53
  • @teenageninjaturtle - Having a publication ≠ notability, or even recognition. The likelihood that your publication will make very little impact at all (along with the vast majority of publications these days) is quite high. I don't mean this personally, I fear this is just the reality. – eykanal Sep 12 '12 at 18:49
  • Good point. One might want to start with posting stuff on arXiv (or a blog) and sending links to people in the field (important: be professional). The feedback (if any) can determine the next step. About (4): that's not generally true (although it unfortunately happens). – Raphael Sep 13 '12 at 7:48

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.