This article describes FAQ FINDER, a natural language question-answering system that uses files of frequently asked questions as its knowledge base. Unlike AI question-answering systems that focus on the generation of new answers, FAQ FINDER retrieves existing ones found in frequently asked question files. Unlike information-retrieval approaches that rely on a purely lexical metric of similarity between query and document, FAQ FINDER uses a semantic knowledge base (WORDNET) to improve its ability to match question and answer. We include results from an evaluation of the system's performance and show that a combination of semantic and statistical techniques works better than any single approach.

This article describes FAQ FINDER, a natural language question-answering system that uses files of frequently asked questions as its knowledge base. Unlike AI question-answering systems that focus on the generation of new answers, FAQ FINDER retrieves existing ones found in frequently asked question files. Unlike information-retrieval approaches that rely on a purely lexical metric of similarity between query and document, FAQ FINDER uses a semantic knowledge base (WORDNET) to improve its ability to match question and answer. We include results from an evaluation of the system's performance and show that a combination of semantic and statistical techniques works better than any single approach.

By typing requests in English, casual users will be able to obtain explicit answers from a large relational database of aircraft flight and maintenance data using a system called PLANES. The design and implementation of this system is described and illustrated with detailed examples of the operation of system components and examples of overall system operation. The language processing portion of the system uses a number of augmented transition networks, each of which matches phrases with a specific meaning, along with context registers (history keepers) and concept case frames; these are used for judging meaningfulness of questions, generating dialogue for clarifying partially understood questions, and resolving ellipsis and pronoun reference problems. Other system components construct a formal query for the relational database, and optimize the order of searching relations. Methods are discussed for handling vague or complex questions and for providing browsing ability.

Although the group at System Development Corp. (Paoli, Pennsylvania), techniques being studied should be widely applicable, we are with each group responsible for certain aspects of system specifically developing a system to understand paragraphlength design. Our groups are jointly responsible for integration of messages about equipment failures, with the aim of the next-generation text-processing system as part of the Defense summarizing each failure and assessing its impact. Advanced Research Projects Agency (DARPA) Strategic Several laboratory prototypes have been constructed for Computing Program (Grishman and Hirschman 1986). We aim to improve on these earlier a small question-answering system that answers simple systems through a combination of two techniques: the use of English queries about a student transcript database This system detailed domain knowledge to verify and complete our linguistic is used for teaching and as a preliminary test bed for analyses and the use of "forgiving" algorithms that some of our linguistic analysis techniques. Participants: Ralph Grishman (faculty); Tomasz Ksiezyk, To guide the development of our system, we selected a Ngo Thank Nhan, Michael Moore, and John Sterling corpus of messages describing the failure of one particular piece of equipment, a starting air compressor.

A scheme for syntax-directed translation that mirrors compositional model-theoretic semantics is discussed. The scheme is the basis for an English translation system called PATR and was used to specify a semantically interesting fragment of English, including such constructs as tense, aspect, modals, and various lexically controlled verb complement structures. PATR was embedded in a question-answering system that replied appropriately to questions requiring the computation of logical entailments.

We have introduced the notion of the closed world assumption for deductive question-answering. This says, in effect, "Every positive statement that you don't know to be true may be assumed false". We have then shown how query evaluation under the closed world assumption reduces to the usual first order proof theoretic approach to query evaluation as applied to atomic queries. Finally, we have shown that consistent Horn data bases remain consistent under the closed world assumption and that definite data bases are consistent with the closed world assumption. ACKNOWLEDGMENT This paper was written with the financial support of the National Research Council of Canada under grant A7642. Much of this research was done while the author was visiting at Bolt, Beranek and Newman, Inc., Cambridge, Mass. I wish to thank Craig Bishop for his careful criticism of an earlier draft of this paper.

This article examines the process of specifying a question-answering help facility in the context of UNIX mail. The specification was based upon experimental expert-user facilitative dialogues. These dialogues were analyzed using a classification scheme developed for the purpose. The scheme provides a metalanguage for describing patterns of intent and rhetorical structure in dialogue. Using the scheme as a tool, common patterns in expert-user dialogue emerged, providing insights into both tutoring strategy and the linguistic forms required to generate help output.

A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity.

Session 6 Logic: II Theorem Proving and STEPS TOWARD AUTOMATIC THEORY FORMATION John Seely Brown Information and Computer Science Department University of California Irvine Irvine, California Abstract This paper describes a theory formation system which can discover a partial axiomization of a data base represented as extensionally defined binary relations.- The system first discovers all possible intensional definitions of each binary relation in terms of the others. It then determines a minimal set of these relations from which the others can be defined. It then attempts to discover all the ways the relations of this minimal set can interact with each other, thus generating a set of inference rules. Although the system was originally designed to explore automatic techniques for theory construction for question-answering systems, it is currently being expanded to function as a symbiotic system to help social scientists explore certain kinds of data bases. Introduction For over a decade researchers in AI have been designing question-answering systems which are capable of deriving "implicit" facts from a sparse data base.

In one major section we discuss the imprecision, the incompleteness, the openendedness, and the uncertainty of people's knowledge. In the other major section we discuss strategies people use to make different types of deductive, negative, and functional inferences, and the way uncertainties combine in these inferences. Keywords Semantics, inference, cognitive processes, natural language processing, human memory, question-answering systems, deduction, analogy 1. Introduction In this paper we will discuss how to represent and process information in a computer in ways that are natural to people. This does not mean doing away completely with representations and procedures which computers have traditionally used, but adding new representations and procedures which they have not used. People often store and communicate imprecise, incomplete, and unquantified information; they often assert truth or falsity in relative terms; and they seldom seem to use rigorous logic in their inferential processes. Because of these conditions, people seem to have an almost infinite information processing capacity, with inference making and problem solving abilities more refined and far more flexible than any existing computer program. How can we study these human capabilities in order to make our machines show similar performance? A combination of approaches is perhaps best. Observation of people's behavior, introspection, some experimentation, protocol analysis, and synthesis of computer programs can all be valuable techniques.