0

We have published a workshop paper last year. Now, we are planning to submit it to a conference. The conference paper roughly have 50% new materials. Basically, the workshop paper studied a 3.5-year data set while the conf version extends it to 4.5-year. The conf version also has a new data set. We also applied several new analysis methods to the data, obtaining new results.

My question is, is 50% new material enough for a new conference publication? Also, can I reuse the analysis methods in the workshop paper? What about reusing some text?

Thank you very much!

Update: Based on the valuable answers and comments from the community, I think a good approach to handle this is issue is to carefully cite our previous work for any content that is originated from it. Therefore, we make a clear distinction between old stuff and new stuff. But a trick thing is anonymity. We probably can only cite our previous work anonymously in the submission, otherwise the reviewers can easy tell the paper is from us...

We will also consult the conference for more information.

3

When in doubt, ask your advisor. Some IEEE venues support iterative publication cycle, workshop, conference, and journal. In robotics, for example, at least 30% of novel material is needed to go from a workshop to conference, and a conference to journal publication. In that case, your new paper is just fine. Other venues, not so much.

In any case, cite the previous paper, and be explicit about what is new in the paper your are submitting.

3

Depends on the specific workshop and conference.

Some workshops (if they are small events at bigger conferences, and if they don't publish the proceedings) don't count as prior publication, so you can use all the material again.

Some workshops are actually more like a regular conference, but use the name workshop. So you should treat it as prior publication.

I suspect from your question that it is more the former. I would suggest asking the papers chair for the conference; if the workshop is related, they are probably familiar with it.

1

You work is published on a CS workshop with proceedings. This is a prior publication. Period.

All CS conferences I know of, require original material. Not 50% original material but 100% new material. This rule will be clearly written in the CFP. For example for WWW conference:

Submissions must represent new and original work. Concurrent submissions are not allowed. Papers that have been published elsewhere, are currently under review, or will be submitted to other meetings or publications while under WWW2015 review should not be submitted to WWW2015.

Or more clearly for SIGMOD:

Every research paper submitted to SIGMOD 2015 must present substantial novel research not described in any prior publication. In this context, a prior publication is (a) a paper of five pages or more presented, or accepted for presentation, at a refereed conference or workshop with proceedings; or (b) an article published, or accepted for publication, in a refereed journal. If a SIGMOD 2015 submission has overlap with a prior publication, the submission must cite the prior publication, along with all other relevant published work, following the guidelines in the Anonymity Requirements for Double-Blind Reviewing section below.

If your plan is to just put the same concepts and contribution, the same proofs and just add new datasets and new experiments and charts, then you should not publish in a CS conference. Instead there are journals that accept extended versions of your conference / workshop papers and you should aim for those.

  • 1
    Note that the SIGMOD snippet specifically mentions what to do when your paper has an overlap with a prior paper (cite it). As such, it seems pretty obvious that some amount of overlap is still ok. Whether just 50% new content is enough is questionable, but I would assume nobody would get worked up if the authors, for instance, had just published the idea earlier without looking at any data at all. Hence I would argue that the spirit of your answer (no prior pub. is ok) is incorrect. – xLeitix Feb 22 '15 at 16:58
  • @xLeitix I do not think that citing one paper, means that in the contribution section you put verbatim the proofs, motivation and just expand the datasets. Also, since a conference is much more prestigious than a workshop, no one would submit an accepted work accepted in a minor workshop to a better conference with just more experiments. In the case of SIGMOD or WWW i can never imagine someone even considering sending an extended version of a workshop paper and in his right mind believing it would get accepted. – Alexandros Feb 22 '15 at 20:02
  • @Alexandros Thank you for your answer. I definitely agree with you that conf is much more prestigious than workshop. However, sometime people might first submit a preliminary version/position paper to a workshop to receive early feedback. Then if the feedback is positive and after significant improvements, a new version could be suitable for a conf. I think for our case, the most important thing is to make substantial new contributions in the conf version and clearly tell the reviewers which portion is from the old one. After all, certain amount of overlapping is probably inevitable. – ZillGate Feb 22 '15 at 21:21

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.