Integration of Domain Problem Solving with Natural Language Dialog: the Missing Axiom Theory 1. Why Should a Computer Speak?

An important problem in developing natural language dialog systems is to computationally specify when and why the system should speak. This paper proposes interruptible theorem proving as a solution. Theorem proving is used to determine when domain goals are complete. Language is used to acquire missing axioms that may be inhibiting proof completion. The paper describes this \missing axiom" theory for the use of language and how it enables the needed dialog processing behaviors to be achieved. The theory is illustrated with a sample dialog segment obtained from actual usage of an implemented dialog system. Performance results of this system based on over 140 dialogs are also given. Designers of fully functional natural language (NL) dialog systems must create an architecture that simultaneously achieves a variety of behaviors necessary for eecient human-machine communication. These behaviors include: the ability to perform problem solving to achieve a target goal. This requires the ability to integrate the required problem solving capabilities within the application domain with the need to interact via NL dialog. the ability to maintain and use information garnered from the dialog. the ability to maintain and use a user model to enable useful verbal exchanges and to inhibit unnecessary ones. the ability to carry out subdialogs to achieve appropriate subgoals and to pass control arbitrarily from one subdialog to another. This paper focuses on the problem of integrating domain reasoning with a general NL dialog facility to enable eeective and timely assistance for task-oriented dialogs, dialogs held in conjunction with the performance of some task that the computer is helping the user to complete. A solution to this integration problem requires an answer to the fundamental question, \Why should a NL dialog system speak?" Interruptible theorem proving is proposed as the answer. In this paradigm theorem proving is used to determine when required domain actions are completed. When the proof cannot be nished due to missing axioms, language may be used in an attempt to acquire the missing axioms. Thus, the purpose for speaking is to provide the information necessary to: (1) obtain the missing axioms; (2) prove completion of the domain actions; and (3) facilitate completion of the overall task. This \missing axiom theory" for the use of language enables the required dialog behaviors to be achieved. Furthermore , interruptible theorem proving minimizes the need for time consuming backtracking in the reasoning process, enabling more eecient dialog processing. This …