What is the "right" rate at which my papers should get accepted (to computer science conferences)? [closed]

Question

In general, when submitting to first tier conferences there is a smaller chance of acceptance but also a better reward.

At one extreme, one may only submit his work to third tier conferences and get nearly all of them accepted at the first shot. Alas, especially for younger researchers, the visibility of these will be very limited and the author might not get recognition for the work at all.

On the other hand, submitting only to first-tier conferences bares a higher risk, especially for PhD candidate looking for a post-doc/position, as it may take several iterations till the paper is accepted and published.

So where should one aim to be in general? Obviously, better papers belong in higher tier conferences, but there may be a goal rate in which your papers should get accepted. If the average number of submissions before accepted is 2/paper (but not all of them are first-tier venues), should one aim higher? If it takes 3-4 submissions but papers end up in first tier venues, should we attempt less competitive venues and publish quicker?

---

Context: I'm a CS PhD student. My supervisor usually insist on sticking to top-tier conferences even if a paper was rejected 2-3 times. I wonder if it is better to lower the bar after a rejection or two to get my results published faster. From another student here I've heard that they usually get papers accepted in the first attempt, but most of her publications are not in a top ranked conference. What is the right balance here?

Answer 1

To a first approximation, promotion depends on top-tier publications and nothing else. If career concerns are important (and they always are) keep submitting to top-tier venues (as long as you believe the work is good), even if it takes several iterations to get in.

There's so much variability in what gets accepted that it's best not to take the outcome of any single set of reviews too seriously. If you start to see a consensus among reviewers as to why they recommend rejection, then of course you will need to address that. But otherwise, don't lose hope and keep submitting.

Context: I'm a social scientist.

Answer 2

I'm sure that this varies a lot by specific discipline, but my perception is that top tier publications really "count," while second tier publications generally don't count for nearly as much during evaluations (i.e. job placement, tenure review). I don't really think there is an equivalence- it's not the case that reviewers will consider two or three second tier publications to be worth one top tier publication, so it's always worth submitting good work to a top conference.

I think the distinction between top, 2nd, and 3rd tier conferences is a little misleading.

I think of top tier conferences as places that publish research that is well done, comprehensive, and broadly applicable. As a result, everyone in the discipline will at least glance at these papers. They also happen to be good papers, because only well done papers can be broadly applicable. The reverse doesn't hold though: merely being a good paper is not enough to be accepted at one of these venues.

Work sent to second tier conferences tends to be much more specific with a much narrower audience. The conferences themselves tend to be very subject-oriented, so people really interested in that specific sub-discipline can disseminate, but the results and the work aren't deep enough that the work would be interesting to someone not familiar with the specific field.

Putting that all together, consider a hiring committee that has to decide between someone with 10 top tier publications and another with 20 second tier publications. The first person has demonstrated that they can produce broadly applicable and interesting work. The second person has demonstrated that they can produce a greater volume of narrow, less general work. When it comes to academics, that first person is going to win every single time.