2

I am an ex-academic (Ph.D + a couple of postdocs) who has been in industry for more than a decade. I work in enterprise analytics and data science.

My current work is fulfilling, but takes up most of my time, and along with my family responsibilities, I don't have the time to work on pure research projects, or to collaborate with an academic team.

I have a couple of serious research ideas in AI and ML, arising from my current work experience plus my previous training and some of the recent results in the literature, that I think are definitely worth pursuing.

They are not just good one liner ideas. I have a rough idea of the steps and experiments needed to implement them (roughly on par with what a serious Ph.D thesis proposal would look like).

As I said, I don't have the time to pursue them myself, but I seriously think they are worth pursuing, and I would like to get credit from coming up with the overall approach, even if can't execute the actual code and simulations myself.

What is an avenue for publishing the ideas and the approaches for investigating them, such that they would get the attention of people who actually have the time to follow through on them, and getting credit for thinking of the approaches?

Does a solid, well substantiated research proposal count as a legitimate contribution?

  • 7
    How about looking for an academic partner to co-supervise students with? – user347489 Dec 18 '19 at 22:59
  • What's your field? In software engineering, there are several possible venues for presenting ideas at halfway stage, including workshops and "new ideas" tracks. Some of them might expect preliminary results, others accept short "position papers". – lighthouse keeper Dec 20 '19 at 8:06
  • @lighthousekeeper I work in Data Science and Machine Learning. – Alex Kinman Dec 20 '19 at 18:38
7

Does a solid, well substantiated research proposal count as a legitimate contribution?

No, it does not. A plan for achieving something, where none of the steps have been completed, is not a significant achievement.

Sometimes it is necessary to do experiments or calculations in order to form a plan. That could be an achievement.

| improve this answer | |
  • 2
    I agree with this, to a point. People get a bit too attached to their ideas sometimes (see all the questions about people afraid their ideas will get stolen). Ideas are relatively cheap, most researchers have lots of them. Putting ideas into practice - securing funding, putting time into research and writing up the results into practice - that's a lot of work. That said, if you have a good idea and won't be putting it into practice yourself, user347489's answer is a good idea about how to have someone else put the idea into practice. – ObscureOwl Dec 19 '19 at 11:01
  • I think that your answer is field dependant, in mathematics for example I would argue that getting the right idea for proving a theorem can be more important than filing in the details, especially if the details follow naturarily. – Nick S Dec 19 '19 at 15:45
  • @ObscureOwl There's always the balance quantity vs. quality. Generally, quantity wins out - when it's the mass of work which should get the credit rather than the neat, but unsubstantiated idea neeings lots of steps to establish. However, occasionally, the key idea is the main contribution without which all that work would lead nowhere (or would not even be undertaken). I have seen several cases of that in my career, so I am not a follower of the "it's the work, not the idea" school of thinking. That being said, there are many "cool" but mediocre ideas around and, yes, these are cheap. – Captain Emacs Dec 19 '19 at 17:12
  • @CaptainEmacs agreed, there's no one answer that's always true for these questions. The contributor of an important idea is, well, a contributor. But someone who contributes an idea then sits back while someone else works several years on it (OP mentioned a PhD project) isn't really a full contributor. That's why I think user347489's answer is a good idea - staying involved as a supervisor allows you to contribute experience > time, and help keep the project on track. Just contributing the idea and then stepping back probably wouldn't be enough on its own to be a full contributor. – ObscureOwl Dec 20 '19 at 8:23
  • 1
    @YemonChoi I am not saying that the details are not important, and often I cringe when people leave the "obvious" steps out... I found mistakes more than once in that part... My comment was more related to the fact that the answer is saying that the idea or plan means nothing without the details, I am trying to say that in mathematics sometimes getting the right idea means a lot, of course completing all the details is also important.... Just to make this crystal clear, if I am struggling to prove a theorem, and someone gives me the key idea for that, I think they deserve to become co-authors – Nick S Dec 22 '19 at 4:21
5

As per Buffy's suggestion I'm turning my comment into an answer.

It seems like you're in a great position to start a collaboration with a university researcher to co-supervise students with them.

Frankly speaking I don't know who your best bet to talk to would be, but reviewing the current literature in the areas of your proposed research is a good idea. Hopefully you're located in a geographical area that allows for an in-person collaboration.

You should also make some research about organizations dedicated to connect academia with industry in whatever country you're located. The cases I know about are the Insight Data Science Fellowships and more broadly all the programs by MITACS in Canada (an maybe even WA state). MITACS programs allow you(r company) to pay half the salary of a graduate student or a postdoc(!) and co-supervise them together with an academic partner in a predetermined project. These internships can be as short as 4 months or constitute a full postdoc.

| improve this answer | |
  • I think this answer gets you closest to "credit", in that a supervisor who does helpful supervising would normally be credited in the student's publication. Whether that's as secondary/supervising author or in acknowledgements should be discussed before the student commits to the project. – ObscureOwl Dec 19 '19 at 11:03
0

What do you mean by credit? You have to be more specific about what is it that you want? Normally researchers publish articles/research outcomes in research journals and are then considered as contributions. If your idea is theoretical, then you can prove it theoretically, try for a publication directly in a theoretical journal or conference (some good venues that allow theoretical research in machine learning are (1) COLT, (2) ALT (3) JMLR. On the contrary, If your idea is experimental, then develop it into a concrete plan of action. Then look for collaborations for carrying out the experimentation. Based on your agreements in collaboration, you can be an author or first author of the publications/patents/startups that result from this. How you collaborate for experimental work is crucial here. You need to convince your team that your idea is worth pursuing and would yield benefits to the collaborators.

| improve this answer | |

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.