Fake research paper detector

If I ever do get to write a PhD, I will have to make sure to run it through this detector (as well covered in the New Scientist). Seriously though, this sounds like a great way to show off the computational linguistics (or more specifically data/text mining) experiments. Hopefully such projects will make the field more visible and more interesting to others.

On the project itself, I wonder how it would deal with papers produced by people for whom English is not their first language.