Deductive Retrieval Mechanisms for State Description Models

Abstract

This paper presents some programming facilities for modeling the semantics of a task domain and for describing the situations that occur in that domain as a task is being carried out. Each such description models a “state” of the task environment, and any given state can be transformed into a new state by the occurrence of an event that alters the environment. Such modeling systems are vital in many AI systems, particularly those that do question – answering and those that do automatic generation and execution monitoring of plans. The modeling mechanisms described are basically extensions and modifications of facilities typically found in AI programming languages such as PLANNER, CONNIVER, and QA4. In particular, we discuss our use of a three valued logic, generator functions to deduce answers to model queries, the saving and maintaining of derived results, and new facilities for modeling state changes produced by the occurrence of events.