2

I sometimes receive a comment from 1-2 of my reviewers saying something similar to "the contributions and novelty of the work/method is limited". Consider it in the context of CS where you develop methods and algorithms which increase the performance of the state-of-the-art in a specific problem/domain.

Of course in every paper, i state my contributions for the reader to understand, but when i receive a general comment like that i get puzzled. Especially when there is no mutual communication between the author and the reviewer to clarify the misunderstandings.

So, let's say, in what typical situation a reviewer might vote for a critism like that?

  • Basically a duplicate of academia.stackexchange.com/q/118335/72855 – Solar Mike Oct 13 '18 at 16:29
  • 1
    Ask the editor or conference chair if they can provide more detail on why it is judged to have limited novelty. Maybe they know something you don't. – Buffy Oct 13 '18 at 16:32
  • @SolarMike: right, i will revise the question in a moment! – Bob Oct 13 '18 at 19:06
  • @Buffy: But considering the high number of submissions i doubt if they make individual contacts with the reviewers to ask them to provide more information regarding their reviews, especially in a timely manner! is it common? – Bob Oct 13 '18 at 19:16
  • Probably not especially common, but no harm to ask. – Buffy Oct 13 '18 at 19:26
2

Possibly related xkcd. Please don’t take offence!

I’ve seen some papers in CS which validly improve upon the state-of-the-art on particular datasets. In fact, you might find these where there is an ongoing annual challenge, or a specific dataset that is used. The numbers are good but they essentially glue two existing methods together. Sometimes there are perfectly novel contributions in the "glue": sticking these two methods together only works if you do it in a particular way, but "glue" contributions are smaller than completely novel "method" contributions and this might elicit such a response from a reviewer. Even if gluing these two methods together is novel.

Usually, in these papers, the results are great (they win the challenge/dataset) and I always wonder why the paper hasn’t been cited more. I think the answer to that is that the citations for the "method" contributions will go to the original papers and the "glue" contributions, if novel, might not transfer to other techniques or fields.

  • Thank you for the comic, i laughed a lot! I understand your point, but in CS, more often they apply established mathematical theories to established base methods to achieve novelty (not just for one dataset or one challenge). And sometimes they called it the glue, and sometimes it is considered chemistry! :D – Bob Oct 13 '18 at 21:02
1

I will take a guess, but it is a guess only.

In CS there are results that seem interesting, even novel, but to someone who knows the field, and its literature, the "interesting" thing is pretty obvious or only a small step away from a known, if obscure, result. If the reviewer's immediate reaction to your main result, is that "I could have easily said that" then you have such a situation.

Of course, the reviewer can be wrong, which isn't as rare as we would like.

For both of these reasons I suggest getting any additional information that you can. Such information can help you revise if necessary, to further distinguish your work from what is known already.

There is another possibility. Perhaps the reviewer has misstated the objection and it should have been "That has limited applicability" rather than "novelty". Again, having more information will help you sort it out.

Unfortunately, in CS, where we depend on conferences, primarily, to present results, the short cycles work against the author in seeking such information.

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.