BibBase is an effort to store BibTeX information as RDF triples. For the data, see: BibBase data.
As of 8 November 2010, there are 6178 publications.
Interesting I suppose but the real question is how to enable researchers using BibTeX to disambiguate their terminology as part of their BibTeX entry?
Has to be as easy as BibTeX and consistent with usage patterns in the communities that use it. If you hope for adoption.
Not hard to imagine a helper application that runs through a set of BibTeX entries and suggest 1998 ACM Computing Classification System or 2010 Mathematics Subject Classification entries. Entries which the author could accept or reject.
Not the fine grained, concept by concept (read subject by subject) analysis of a document that I would like to see, but it’s a start.