-5

I recently defended my master's thesis in artificial intelligence. Among other things, I invented a new algorithm that I would like to publish. The domain is natural language processing, with papers viewable for free on ACL Anthology.

For backstory: I developed my thesis almost entirely by myself. I only saw the professor who commissioned the proposal 4 times (which includes the final defence) as she resides in a different country, and although I had weekly meetings with my mentor (one of her PhD students), he was always behind on my research and hence he mainly helped with e.g. stylistic choices. I don't blame him for not having made deep contributions, since (1) it wasn't his thesis and (2) he wasn't as deep into the literature as I was throughout the year.

Now, although I included numerical experiments in my thesis, he correctly pointed out that my algorithm should be tested as part of a computationally heavy model using industry-standard metrics if I were to publish it in a paper. Being a PhD student, he has access to computational resources I don't as a mere alumnus. So, it made sense to let him collaborate on this paper.

We started a shared LaTeX document a while back, and unsurprisingly, the vast majority of all the text (intro, related work, theory, algorithm, discussion ...) will have been written by me, since I developed my thesis and hence I know both the algorithm and the literature thread that led up to it best. I also wrote all the code for my algorithm, because I needed it in my thesis. His contribution will be general remarks about the text, and providing me with the computational resources and interfaces to run large-scale experiments, so I expect that he will likely be writing that section.

I don't blame him at all for this distribution of the work; it is natural given the backstory. He's not freeloading, because his access to more, bigger experiments is a valuable addition.


My issue is that I want to ensure that I, a fresh graduate with zero publications, get maximal public recognition for the year-long independent work I did to reach this literary thread and this novel algorithm. If this paper is ever cited, it will be cited as Mew & Him (2023), and to be honest, it breaks my heart just seeing it that way. It is my brainchild. I rarely see a paper cited for its numerical experiments rather than the theoretical idea presented by it, and hence I lament that such a citation only seems to give me half the credit. That does feel like freeloading, even though it isn't his intention nor his fault that citations work that way. It feels like I'm selling half my credit for a single (useful and hopefully interesting) section.

He and I are are amicable, and I'd like to keep it that way. However, since my thesis was quite large (170+ pages), more than one paper can be distilled from it, and he has already made it clear that he would want co-authorship on all such papers since it is supposedly customary to give one's thesis advisor co-authorship, even though him and I both know that the content of the thesis was entirely my own doing. Whilst I respect this -- and furthermore, he did have to sit through weekly meetings with me mostly failing to articulate mathematical ideas orally -- I feel like I'm being partially robbed of making a name for myself for a unique contribution to the field. For him it seems like a closed case, and I'm not sure how to handle peddling it back. He finds my ideas valuable and has interest in them, but being interested and receiving recognition (as one of the 2 authors) are quite opposed. What should I do?

If he becomes co-author of these papers, then at the very least, I would like to somehow indicate to the reader that I was responsible for the idea and he was responsible for the experiments. Otherwise, since he already has publications, it might even seem to the reader like I am the novice freeloading off the master, and that the order of authors should secretly be reversed. How can I handle this as neutrally as possible to the reader, and communicate my desire about this to my mentor?


Perhaps I'm just insecure about lack of recognition from past experiences, or perhaps I'm insecure about possibly being a one-trick pony... but that pony still deserves proper credit for its one trick.

There is a similar thread on the site here, but quite a large emphasis is placed there in the answers on the fact that the paper has already been submitted, and also, the author is a PhD student publishing with higher-ups. In my case, the first paper has largely been written -- except for the experiments -- but has not been submitted for review, and I'm no longer a student (but might start a PhD in the near future), making my mentor moreso my colleague than my boss -- but obviously, he did have power over me when he was in charge of overseeing my thesis backed by the professor.


Edit: Apparently people are missing the part where I state that I already did my own numerical experiments to verify my work. Mea culpa, I shouldn't have made the post so long. What I contributed is the literary background, the idea, the mathematical formalisation, the formal algorithm, the practical algorithm, and the small-scale testing.

It might also be relevant to stress that the small-scale testing was intrinsic, i.e. it measured the effectiveness of the algorithm using metrics made to do exactly that. The large-scale testing would be extrinsic, i.e. measuring its effectiveness using metrics to evaluate the large-scale system, merely as a proxy for my one component. So yes, I already know my idea works, formally and empirically. The extra experiments would be just that: extra.

7
  • 2
    If you do not know whose idea it was to write the proposal, you cannot possibly know whether this other person should (ethically, not just by convention) be on the paper or not. It is difficult (and thus valuable) to identify problems that could be solved. Put bluntly, you executed someone else's plan, as almost every student does initially.
    – user_phys
    Aug 30, 2023 at 3:55
  • 9
    In my field it is very uncommon to have single author papers and even less common to have single author papers from students. I do not think that being the sole author would in any way increase your cerdibility or standing, rather people might wonder why a student did publish research without any of their supervisors (and many of the scenarios that spring to mind are not very positive). I feel like you have a skewed view of how publishing with co-authors is perceived. And from what it sounds like (although hard to tell for sure) the contributions of the PhD student do merit co-authorship.
    – Sursula
    Aug 30, 2023 at 6:38
  • You write that “it is supposedly customary to give one's thesis advisor co-authorship.” My advisor didn’t request—and likely would have refused—credit for a paper I published from a chapter of my dissertation in comp. sci. Granted, he’s a really good guy. Aug 30, 2023 at 12:10
  • 1
    I've seen several posts where the author thinks too highly of coming up with an idea, say, for an experiment. Everyone should get credit for what they do, be it writing, proofreading, theorizing, experiential setup or design, or just getting lab work. Give credit precisely as it is due. Often journal have a section for detailing each author's specific role.
    – jdods
    Sep 2, 2023 at 17:04
  • 2
    I'll give you the chaotic-neutral advice: add a 3rd author so it gets cited as "Mew et al" (yes, I'm joking. Although technically it'd work)
    – Pronte
    Sep 3, 2023 at 16:33

5 Answers 5

9

tl;dr:

Remove all contributions of your co-author from your paper. What is left of your paper? no correct mathematical formalization of your idea, no testing of the algorithm, just an idea and an algorithm applying it. Treat your algorithm as an experimental apparatus. Without a rigorous formal mathematical description, you are bound to err (see this dramatic example of how dangerous is an idea alone).


My issue is that I want to ensure that I, a fresh graduate with zero publications, get maximal public recognition for the year-long independent work I did to reach this literary thread and this novel algorithm.

You can show it at due time: conference and when you give a talk. But please say explicitly "I had this idea and my co-author did have to sit through weekly meetings with me mostly failing to articulate mathematical ideas orally". It will be then clear who had the idea and who formalized it (please note: I am being a bit tongue-in-cheek).

If this paper is ever cited, it will be cited as Mew & Him (2023), and to be honest, it breaks my heart just seeing it that way. It is my brainchild. I rarely see a paper cited for its numerical experiments rather than the theoretical idea presented by it, and hence I lament that such a citation only seems to give me half the credit.

Without the numerical experiment, your paper will be perceived like "yes, crazy idea, how did this unknown crank published it on a peer-reviewed journal without presenting serious testing, let's wait for someone to perform proper tests on it" which means no one will test it, no one will prove it, no one will cite it, maybe 2-3 years down the line some master thesis will prove it and publish and cite your precious idea.

With the numerical experiment it will be perceived like "yes, crazy good idea, this unknown person published it on a peer-reviewed journal with a good set tests, relevant, I may build on it with citation". The work can then enter in the flow of citations of your field.

The numerical large testing experiment is then necessary. Would you have done without your advisor? no. Therefore it was their idea to perform them and they deserve recognition for that, so a co-authorship seems to me relevant.

Having an idea is a necessary, but not sufficient, requirement to publish a paper.

6
  • I don't understand where you are getting the impression that I didn't formalise my idea. Communicating mathematics orally is extremely difficult because mathematics revolves around manipulating written symbols. My mentor did not formalise my ideas, I did all of that work myself; like I said, his only contribution in a paper describing, formalising and algorithmising the idea is providing access to indirect empirical verification. I already provided the direct empirical verification in my thesis because that was possible on a small scale, and it checks out.
    – Mew
    Sep 5, 2023 at 15:42
  • For your TL;DR to be accurate, it would have to read: "Remove all contributions of your co-author from your paper. What is left of your paper? An idea with backing in literature, a correct mathematical formalisation of your idea, an algorithmic implementation, results from small-scale testing of the algorithm, but no large-scale high-resource tests." This is almost the polar opposite of how you wrote it.
    – Mew
    Sep 5, 2023 at 15:45
  • dear OP, you wrote this: he did have to sit through weekly meetings with me mostly failing to articulate mathematical ideas orally. Me and everyone that thought my answer is relevant understood that you were failing in articulate mathematical ideas orally. If you meant differently, then you failed in communicating it. I understood wrong? ok, my fault. Many people understood it wrongly, ok, your fault. This is only one anedoctal point, but based on your question I invite you to think carefully about how you communicate with people:.d
    – EarlGrey
    Sep 6, 2023 at 7:20
  • I didn't mean differently. I meant it exactly as I wrote it, hence orally. Nowhere did I say that I failed to communicate them in writing. I'm not a mathematician, but I have an inkling that they too would rather have a piece of paper or a blackboard at hand to explain their ideas than to handwave, resulting in their peers failing to understand their in principle correct train of thought.
    – Mew
    Sep 7, 2023 at 21:12
  • @mew why were you torturing your advisor with these weekly meeting in their format, then? Were you prohibited from bringing a paper and a piece of paper? Anyway, it looks like you can publish the paper on your own, go on and be happy!
    – EarlGrey
    Sep 8, 2023 at 7:12
7

If you want a career in academia, I think you are making a serious mistake by focusing too much on a single paper, especially if you reject the contributions of a collaborator.

Long term, you are probably better off being generous than selfish about such things as they can lead to the next paper and the one after that and additional people to share ideas with, extending your field and boosting your career.

How often does a paper come along that a career is "made" based on a single paper by a single author? Even Einstein had a lot of collaborators and a difficult time starting his career. It is easy to look back on AE and forget how difficult it was for him.

Also, note that "ideas" aren't enough. It is in the verification of those ideas that value is assigned. Often said that "ideas are free", though some would disagree.

2
  • wrt to Einstein, it is worthwhile to note that some autorship of his first papers is still debated (was the wife working with him?) and plenty of people had the same ideas as Einstein: en.wikipedia.org/wiki/… The paper from Einstein however. although it shares the same idea as those previous paper, is the most complete and coherent formal description, maybe he missed some citations, but for sure his work is the most complete and remembered. OP may rethink about uniqueness of ideas ...
    – EarlGrey
    Aug 30, 2023 at 13:16
  • (1) It's not a single paper, it's multiple papers. (2) It is a luxury to be generous. It's easier said than done when 100% of your net worth is in the balance. (3) I'm not asking to make a career from a single paper. I'm asking to capitalise the most possible off of it in order to maximise the opportunities to make a career afterwards. (4) As stated in my post and re-iterated in the edit, I already explicitly verified the idea.
    – Mew
    Sep 5, 2023 at 16:01
6

There is such a thing as the contributions section where the contribution of each other is specified. In some journals such a section is already mandatory, exactly for the purpose to provide the reader with the information which author was responsible for what.

So asking for such a section to be added to your paper is not only reasonable, but also not that unusual.

4

I also did research in NLP for quite a number of years, and at least in NLP specifically, single author paper is rare, and, to cater your worry, in a two-author paper people never think "oh, the first author must be freeloading off the more capable second author".

In NLP, so far I've seen only long-term professors who have published as solo author. It's completely normal and extremely common to have two-author paper in NLP. And when that paper is good, so far I've seen that the NLP community is giving credits to the first author.

tl;dr It's fine to publish a two-author paper in NLP, first author still gets the recognition.

0

A professor in AI and ML here with years of experience evaluating CVs, people, reviewing papers, editing journals etc.

It is extremely rare to see single authored papers these days, unless the author stems from a theory setting. I know even within ML different corners treat authorship differently : for example in theory all others must have contributed to the idea, the development, independently if they have funded the people involved to be called co authors whereas other corners will be more generous in offering co-authorship.

Nobody will think less of you for having two authors or even three authors. To the contrary it shows your willingness to collaborate, to offer credit where credit even if tiny is due. Science is not an "I" business anymore but a team affair. Learn how to develop a family around you that can help you in the future.

Now, I am surprised that you don't mention the MSc advisor at all. Have you consulted with them? Are they happy to not be a coauthor?

In my book not adding the PhD student who helped is not fair. If you want to be truly open as to the contributions of each author you can add an author contribution paragraph or footnote.

Unless you have a very exotic last name, I wouldn't worry about how it appears in citations. In some referencing styles, it would come as a plain number.

1
  • Welcome to Academia.SE! I do think of myself as theory-oriented, hence why I'd need external support to run experiments on a supercomputer rather than my small-scale verifications. With respect to single-author papers, this one is quite recent yet renowned in NLP, cited as Kudo (2018). To answer your question: no, I have not contacted the professor under whom I did the thesis. She is acknowledged at the end for assessing the thesis from which the paper came, but she was uninvolved in any of the research (since I only saw her 3 times for the whole year).
    – Mew
    Sep 13, 2023 at 13:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .