Bergmair, Richard. 2006. “Closed Domain Question Answering Using Fuzzy Semantics.” M.Phil. thesis, Cambridge, England: University of Cambridge.
In the present report we give a thorough exposition of our first steps towards a theory of fuzzy semantics and towards the development of a closed domain question answering system in the form of a natural language database interface that produces result sets with records ranked according to the degree to which they fulfill our intuitions about vague expressions in natural language, and vague adjectives in particular.
We outline our ordering-based approach to semantics and introduce some of the issues involved in modelling vagueness. We show how fuzzy sets can be used as intermediate semantic representations of vague expressions. From the family of possible fuzzy logics we pin down which one best fits our modelling needs in fuzzy semantics.
We then describe the overall design of a controlled experiment involving human subjects, the software infrastructure necessary for administrating it and the statistical analyses required to draw conclusions from the data about the adequacy of our model of fuzzy semantics. We will also discuss the results from a small-scale preliminary instantiation of this experiment.
Finally we show how we could put those theoretic insights to use in a working natural language interface to a database that produces rankings of “small cities” or “rainy cities near San Francisco”.
==> director’s cut
[PDF]
==> as
submitted [PDF]
==> slides used at viva
[PDF]
==> slides
used at Friday seminar [PDF]
==>
at 23C3
==>
video at 23C3 [PDF]
==> 23C3 slides
[PDF]
==>
at ResearchGate
==>
at archive.org