DL and Grammar Parsing Analogies

A blog post that I intend to write will sketch my observations from the vantage points of ontologies and natural language processing.

Most of my work in using ontologies for knowledge representation and reasoning has used various description logic languages, and most of that has been in developing and using a reasoner for EL++, one of the OWL profiles. And most of my work in NLP has used a Feature Logic reasoner which I developed.

There are a few papers that I have read that compare DLs and FLs and a few papers that explore using both in some interesting ways.  I will address some observations I have about how each discipline can inform the other, and how I have found some useful techniques for applying them in tandem. A few topics:

  • TBox akin to fully compiled grammar
  • ABox akin to assertions of word types and positions
  • DL complete and correct reasoning
  • Grammar incomplete and speculative reasoning
  • Priming and most probable parse selection
  • How do DLs inform grammar processing?
  • How does grammar processing inform DL technology?