An Efficient Compositional Semantics for Natural-Language Database Queries with Arbitrarily-Nested Quantification and Negation

  • Authors:
  • Richard A. Frost;Pierre Boulos

  • Affiliations:
  • -;-

  • Venue:
  • AI '02 Proceedings of the 15th Conference of the Canadian Society for Computational Studies of Intelligence on Advances in Artificial Intelligence
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

A novel and efficient implementation of a compositional semantics for a small natural-language query processor has been developed. The approach is based on a set-theoretic version of Montague Semantics in which sets that are constructed as part of denotations of negative constructs are represented by enumerating the members of their complements with respect to the universe of discourse. The semantics accommodates arbitrarily-nested quantifiers and various forms of negation, nouns, transitive and intransitive verbs, conjunction, and disjunction. Queries containing the word "is" and passive verb constructs can be evaluated. However, the approach for these two constructs is ad hoc and inefficient. The approach has been implemented in a syntax-directed evaluator, constructed as an executable specification of an attribute grammar, with a user-independent speech-recognition front-end.