Comment 265

next comment

Tags: * poor ARR meta/review quality * need new reviewing criteria

"The quality of reviewers is very low in ""prestigious"" acl venues. Mostly anyone (even master's students) 
seem to qualify as reviewers and this can be seen in the reviews. If you want to get a paper accepted 
in an ACL venue, it's best to do some latest stuff in neural models and propose state of the art results. 
If you don't use latest in neural models, the inexperienced reviewers will complain. If you do error 
analysis, they will tell you your results weren't too good after all. If you deal with languages that 
have limited resources, they will tell you that this has been done before for non-English languages 
(because all of them have the same resources and are in all ways identical anyways)... In short, it's 
easier to get a machine learning garbage paper accepted than an insightful paper that challenges the 
current state in the field. "