During my research on MOOC's I started to work on a system that is able to automatically estimate the quality of a review based on the underlying text. However testing such a system is not easy as one needs a data set of written texts, some free text reviews for them, and a quality score for the review. Does anyone know whether such a data set exist`s? Would anyone dare to put her own paper and the received reviews for the paper into such a pool.
Maybe something like this could be used: http://publications.copernicus.org/services/public_peer_review.html
Hi Bastian,
thx for the answer. This is actually quite close to what I am looking for. Although there seem to be quite few comments for the paper. But I will definitely have a closer look.
Nips now publishes the reviews of accepted papers, if that helps...
Http://papers.nips.cc/book/advances-in-neural-information-processing-systems-26-2013
Thx Vladimir. Another interesting source. It seems that only the final review is public is that true?
Hi Markus,
I don't know about a central resource for published reviews, but the ICLR conference is asking for public reviews and discussion of papers. They did it this year and want to keep that publishing model:
https://sites.google.com/site/representationlearning2013/program-details/publication-model
Papers and reviews can be found at:
http://openreview.net/venue/iclr2013 and
http://openreview.net/venue/iclr2014
Maybe openreview.net could become a resource that is useful for you, but it seems pretty new and is just starting to host the review process for more conferences.
Perhaps EasyChair maintains a database with reviews of all the conferences that have used it to organize the review process.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence