> I suspect most of them
    > would have gotten away
    > with it if they changed
    > the title a bit

Cosine similarity over a corpus of published papers to find very similar ones would be an interesting exercise