Tags: * ARR excludes interdisciplinary research * against ARR * increase ACL paper length
"I am working on interdisciplinary field between NLP and other areas of computer science. I feel that
ARR has just kicked me out of NLP. Practically, I believe ARR has had a strong side effect of eliminating
papers in the interdisciplinary field. In fact, after I left *ACL conferences because of ARR last year,
my papers were accepted in top conferences other than NLP, in which no ""rolling review"" is conducted.
This strengthened my belief. Understanding interdisciplinary work requires open-mindedness to understand
different practices and different prerequisite knowledge other than NLP. Some reviewers refuse to understand
it in the first place. So, shuffling reviewers until all reviewers are open-minded reviewers is essential
for accepting interdisciplinary works, which is against the concept of ""rolling review""s. Moreover,
the ACL paper format has considerably less space than other conference formats, such as the ACM format.
Requiring brevity has a strong effect that prefers mainstream works and native English speakers. I feel
that submitting papers to *ACL is now merely a guessing game as to what the NLP reviewers know, which
is not essential nor academic at all. Simply speaking, ARR is too complicated and excluding interdisciplinary
works. Just return to a system where reviews and meta-reviews are written and used for just one conference."