Let’s stop reviewing for publishers that profit from research

What’s the deal with how we review papers for venues (like conferences and journals) for free and then they go on to sell and restrict access to them? How about let’s only review for venues that freely distribute papers and stop reviewing for those that restrict access? We can also stop reviewing for those that charge high publishing charges, I believe over $100 per submission is unacceptable.

We have the power to put an end to closed access research. By only reviewing for venues that freely distribute papers, we will ensure they have the best publications and become the premier venues. It will then become in everyone’s best interest to publish in venues with freely accessible papers. 

This attitude has transformed machine learning publishing after the editors of “Machine Learning Journal” (MLJ) quit to create JMLR. Without the ability to referee papers with good reviewers, a journal cannot function or will be forced to select from lower quality reviewers which is a negative feedback loop. In many fields (and growing in machine learning again) it appears that those in power of the venues which are thought of as “required for academic career advancement” do not share the ideal that all researchers should be able to access research whenever they want even if they are poor.

It is urgent that we stop making the issue worse. The current system creates disadvantaged researchers, why do we want that? During my entire academic career (from community college to PhD) I have been blocked by paywalls and forced to use libgen and sci-hub. Every new paper that is published in a closed journal is another paywall that a future researcher will see. We may always need to resort to piracy sites for past papers but we are in the position to prevent future papers from being locked away.

So what can we do? In 2014 I started sending this letter in response to each request to review from one of these venues. Feel free to use it! It is important to let the editors know why you are not reviewing so that they can rethink their own participation. Reviewers are a critical part of a conference or journal’s operation. Striking is a powerful way to push for change!

I’m sorry but I no longer review for journals which limit access to research in order to sell freely submitted research publications or charge an open access processing charge more than $100. I do not appreciate a business making a profit off my effort as a reviewer by charging other researchers to read articles or publish them. I instead urge your journal to follow a format similar to:

– The Journal of Machine Learning Research (JMLR)
– Proceedings of the Very Large Data Bases (PVLDB)
– Neural Information Processing Systems (NeurIPS)
– International Conference on Machine Learning (ICML)
– International Conference on Learning Representations (ICLR)
– Medical Imaging with Deep Learning (MIDL)
– Journal of Machine Learning for Biomedical Imaging (MELBA)

These two articles echo my reasons for not supporting this model of publishing (I also do not support “pay to publish” open access journals):

http://jmlr.org/statement.html

http://www-cs-faculty.stanford.edu/~uno/joalet.pdf

Read more about this statement here: https://josephpcohen.com/w/reviewerstrike/