What exactly is (in your opinion) the difference between

  • workshop papers,
  • conference papers, and
  • journal papers in NLP?

I mean, it's clear that the latter are for more mature and more relevant work than the former. Any other characterizations?

What about a publication that tackles an interesting problem with an uninteresting, unoriginal hack and achieves great results? Definitely workshop paper?

What if there is a revolutionary new idea and method, but it hasn't been evaluated much empirically yet? Could be a conference paper?

What about a tutorial-style paper that introduces a method thats' known in machine learning to the NLP community and describes some applications to typical NLP problems?

asked Jul 23 '10 at 11:33

Frank's gravatar image


One Answer:

There are nominal differences between a workshop and a conference paper. A workshop paper isn't technically a publication and is typically meant to represent work-in-progress. Nowadays, workshops are typically very topically-focused and non necessarily any less quality than a conference paper. That being said, I think its safe to say workshop papers are generally regarded as less prestigious than conference papers (at least in NLP).

Journals play a weird role in NLP. Good work which just needs a lot more room for exposition takes good advantage of journals. Generally, they are nowhere near as read as conference proceedings. Because of the length of time it takes to get from submission to publication, people generally will target really good work at conferences, and then may submit a journal version.

answered Jul 23 '10 at 11:42

aria42's gravatar image


edited Jul 24 '10 at 10:39

Alexandre%20Passos's gravatar image

Alexandre Passos ♦

Your answer
toggle preview

powered by OSQA

User submitted content is under Creative Commons: Attribution - Share Alike; Other things copyright (C) 2010, MetaOptimize LLC.