Issues in the development of natural language front-ends

ABSTRACT

This paper will discuss some issues we believe to be important to the design of a natural language front-end. These are divided into three categories: conceptual coverage, linguistic coverage, and implementation issues. The section on conceptual coverage discusses the use of a domain expert, which understands what the user is saying even though the system to which the front-end is interfaced might not be able to properly do what the user wants. The section on linguistic coverage discusses attempts to allow a natural language interface to handle natural, interactive human communication. Two solutions are explored: First, the design of a robust natural-language-understanding system, composed of many experts that know about some aspect of the organization of language, is considered; second, because the design of a robust system is a large task, the intermediate goal of limiting the vocabulary and constructions that can be used while retaining all the user-oriented benefits of natural language is considered. The implementation issues considered are the design of a system in which the grammar and the domain of discourse can be easily extended and which can be used for more than one domain without extensive rewrite.

REFERENCES

• 1: G.M. Birtwhistle , O.J. Dahl , B. Myhrhaug , K. Nygaard, Simula Begin, Chartwell-Bratt Ltd, 1979
• 2: Chapanis, A., Parrish, R. N., Ochsman, R. B. and Weeks, G. D. Studies in interactive communication: II. The effects of four communication modes on the linguistic performance of teams during cooperative problem-solving. Human Factors, 1977, 19, 101--126.
• 3: Hewitt, Carl. Viewing control structures as patterns of passing messages. A. I. Memo 410. Cambridge, MA: MIT AI Laboratory, 1976.
• 4: T. P. Kehler , R. C. Woods, ATN grammar modeling in applied linguistics, Proceedings of the 18th annual meeting on Association for Computational Linguistics, June 19-22, 1980, Philadelphia, Pennsylvania [doi>10.3115/981436.981472]
• 5: Kelly, M. J. and Chapanis, A. Limited vocabulary natural language dialogue. International Journal of Man-Machine Studies, 1977, 9, 479--501.
• 6: Michaelis, P. R. Cooperative problem solving by like- and mixed-sex teams in a teletypewriter mode with unlimited, self-limited, introduced and anonymous conditions. JSAS Catalog of Selected Documents in Psychology, 1980, 10, 35--36. (Ms. No. 2066)
• 7: Brian Phillips , James Hendler, The impatient tutor: an integrated language understanding system, Proceedings of the 8th conference on Computational linguistics, September 30-October 04, 1980, Tokyo, Japan [doi>10.3115/990174.990261]
• 8: Roger C. Schank, Conceptual Information Processing, Elsevier Science Inc., New York, NY, 1975
• 9: Schank, Roger C. and Kolodner, Janet. Retrieving Information from an Episodic Memory, or, Why Computers' Memories Should Be More Like Peoples. Yale University Depart, of Computer Science Research Rpt. 159 1979.
• 10: Norman K. Sondheimer , Ralph M. Weischedel, A rule-based approach to ill-formed input, Proceedings of the 8th conference on Computational linguistics, September 30-October 04, 1980, Tokyo, Japan [doi>10.3115/990174.990183]
• 11: Harry Ralph Tennant, Evaluation of natural language processors, University of Illinois at Urbana-Champaign, Champaign, IL, 1981
• 12: Bozena Henisz Thompson, Linguistic analysis of natural language communication with computers, Proceedings of the 8th conference on Computational linguistics, September 30-October 04, 1980, Tokyo, Japan [doi>10.3115/990174.990206]
• 13: David L. Waltz, An English language question answering system for a large relational database, Communications of the ACM, v.21 n.7, p.526-539, July 1978 [doi>10.1145/359545.359550]
• 14: Ralph M. Weischedel , John E. Black, Responding intelligently to unparsable inputs, Computational Linguistics, v.6 n.2, p.97-109, April-June 1980
• 15: Wilks, Y. A preferential, pattern-seeking, semantics for natural language inference. Artificial Intelligence 6, 1975, 53--74.