Cooperative Responses in Deductive Databases

Annie Gal

PhD Thesis
Computer Science Department
University of Maryland
College Park

The complete paper is available in:

Abstract

This thesis addresses the problem of full utilization of the semantics already present in a database, and particularly, integrity constraints, to respond more cooperatively to a questioner. This thesis provides a new use for database integrity constraints, primarily used for maintaining database consistency during updates.

We specify a general apprach to the problem. We explain how to collect information relevant to the user's query, so as to give the user a cooperative response. Then, we describe rules to select from among the collected information, to provide a clear and concise informative answer to the questioner. A general set of heuristic rules are offered for that purpose. An implementation of this approach in the form of a natural language interface is provided.

Our approach to generating cooperative responses is domain-independent, although it uses domain-specific information (contained in the form of integrity constraints, the user's query and the database itself). The method utilizes knowledge already present in a database and can be applied to any deductive database (and also any relational database, since deductive databases are generalizations of relational databases).

Bibliography