4

A similar question was asked in Having a lot of papers published in unimportant conferences VS very few in good ones?, but my question is only about reputation, so the answers there do not fit my question.

Background

I am working in the field of Machine Learning and for my PhD I need to publish about 3 papers. Before I started my PhD I have been working in industry and am still working in industry half-time (just to earn money, it has nothing to do with my PhD). I do not plan to stay in academia after I finished, but plan to still have connections with academia/research and plan to continue to publish. For my PhD I am only allowed to publish on A1 or A2 conferences (according to Qualis). So for example "SIGKDD" or "NIPS" would be a A1 conference and "ECML-PKDD" or "SIAM DM" is an A2 conference - so they all need to be pretty good, I guess.

Main Question

I would like to know whether what is better regarding reputation: Publish fewer papers on better conferences or more paper on worse conferences. How about better/worse journals?

If you can, please answer for a) the academic world, and b) for the business world.

Clarification

I suspect that publishing on a conference that is worse than A2 is a waste of research time and one might be better of just putting a bit more extra work into the research to publish in at least on a A2 conference.

Additional Question 1: Is that assumption right?

I have only submitted to A2 conferences, so, I do not know how much more work it would be to be able to be accepted to an A1 conference. I guess in both cases you need to do a similar amount of base research work and then some more work which depends on whether it will be A1 or A2 worthy. I would suspect that one might be able to publish twice as many papers (as in "would need half as much research and writing time") if one would only submit to A2 if one would only submit to A1 conferences.

Additional Question 2: Does this rough estimate make sense?

Regarding conference vs journal: We publish mostly in conferences because of the shorter review process. However, after my PhD this might be not so important to me, but instead it will be more important how much work I have to put into it in total (not the length of time from start to completion). It takes me quite a lot of work to shorten all my texts so it fits the page limit. Also, a lot of the negative feedback I get from reviewers is that something was missing, that I had to leave out, because of the page limit. Rejections, of course, cost work (because you need to rewrite text for a different format - we are not counting time I need to additionally spend, because my research not good enough, of course). I think that it might actually less work to write for a journal because of these two points, but I am not sure.

Additional Question 3: Does this makes sense?

4
  • How about "all of the above"? Lots of papers with a few in the best conferences, assuming you can get accepted. Don't let the perfect be the enemy of the good. – Buffy Jun 6 '20 at 13:55
  • 1
    "Business world" is too broad. For many positions papers mean little if anything. Or do you mean industrial research? – Buffy Jun 6 '20 at 13:56
  • 1
    @Buffy: Sort of "industrial research as a freelancer researcher data scientist" or "as a founder of a startup doing innovational stuff". or "innovative speaker on AI". This kind of area. If you say, it all does not matter - than I that might also be an answer. – Make42 Jun 6 '20 at 13:59
  • 1
    "all of the above": Well, I would want to have a good return of investment, where "invest" is time and "return" is the reputation. If you say, that a mixture would have the best ROI, that might be an answer. – Make42 Jun 6 '20 at 14:00
2

I'll suggest that trying to game this out in advance is a mistake. You will miss opportunities and focus on the wrong thing. Do the research. Write the papers and only then decide on an appropriate venue. If you think that a bit of extra effort on a certain paper might help it be accepted in a better conference, then do that. But otherwise send it to an appropriate conference. If you wait a year, trying to improve it, you will get scooped in a hot field.

You don't want to be the person that only ever did one significant thing and was otherwise unknown in academia.

First the research, then the papers, then decisions. That is different from not having a goal, of course.

But being present in lots of conferences also lets you build your circle of collaborators.

4
  • 1
    My particular research interest is everything but hot... I research unsupervised learning foundations and there are very few (good) papers published compared e.g. to supervised neural networks. – Make42 Jun 6 '20 at 15:48
  • Sure, I understand what you mean. But let's say I finished by research milestone and I could publish to a very good venue with a deadline in 1 month or a top venue in 4 months (and I would also need those extra 3 months to improve it to a point to be able to put it into a A1 venue). – Make42 Jun 6 '20 at 15:51
  • Three months isn't a terribly long time, but I hope you don't think that either path is sure to result in acceptance. I don't know how much experience you have at this, but some top conferences are pretty competitive. And nothing prevents you from submitting to one, but continuing to work on the ideas, either to improve the given paper or to write a follow on. – Buffy Jun 6 '20 at 17:04
  • I have have submitted about 5-6 times and got 2 acceptances in a good (A2), but not a top conference (A1). So I am not sure what it takes to be accepted to a top conference in comparison to a good conference. My supervisor said it is more difficult, but could not really give any more details. Maybe it is difficult to put into words. Can you enlighten me on this regard? – Make42 Jun 6 '20 at 18:34
0

In general my unequivocal advice is: focus only on the top to very good venues/conferences. Do not submit to "worse" places, unless (and only in that case) you simply do not expect to be able to get accepted in top places with the minimal amount of papers to sustain your career.

Example: The following case 1 will advance your career much better than case 2, but 2 is still better than 3:

Case 1: Publish 4 papers in top to very good venues during PhD.

Case 2: Publish 12 papers in "non top or not even good" places during PhD.

Case 3: 1 paper only in top place during PhD.

In case 1 you're going to get top postdoc positions with some high chance. In case 2 you are sure not to get any top postdoc positions. But may get something decent. In case 3 you may be in trouble not getting any postdoc/job offers at all. Hence, 2 is better than 3.

Comment: I am talking only about academia. I don't know much about industry in that respect.

4
  • Thank you for your assessment. Regarding my PhD, as I wrote, I am required to publish in top to very good venues, so I have not say in the matter. What you wrote supports my assumptions for "additional question 1". My question has never been whether I should publish "worse than 'top to very good'", but whether it should be "top" (A1) or whether "very good" (A2) is enough. I suspect that A1 requires quite even more work and my question is basically, whether that is worth it - compared to A2. What do you think? – Make42 Jun 6 '20 at 15:47
  • Interesting ranking. However I cannot endorse it. It has mistakes that make it invaluable to me. Also, it has clear inconsistencies: the same conference is ranked differently! E.g., A, A1, or A, B. Etc. – Dilworth Jun 6 '20 at 21:17
  • No, you are making the same mistake in reading the ranking as I did at first: The webpage conferenceranks.com is only reporting the rankings of different institutes: "ERA (2010)", "Qualis (2012)", and "MSAR (2014)". These three have different ranking schemas. ERA has only A,B,C, while Qualis is more detailed, like A1,A2,... When checking out the "Rank", you need to also check out the "Source". – Make42 Jun 8 '20 at 9:37
  • I see now. You are correct. it is not a bad ranking system. But I wish they would include more ranking authorities, like CORE. – Dilworth Jun 8 '20 at 13:54

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.