A Survey of Interfaces to Data Base and Expert Systems.

Moerdler, Galina Datskovsky

The problem of natural language recognition has been studied by many researchers since the early 1960‘s. In particular, researchers are concerned with building interfaces to existing systems, such as data base and expert systems. Such interfaces focus on easing communications between the underlying systems and the user. In this paper we survey various interfaces to both expert and data base systems. Natural language interfaces to data base systems have been quite successful. In contrast, natural language interfaces for expert systems are still in the initial stages of design. One mode of communication between an expert system and a user is via a menu interface. Such an interface presents many limitations [Datskovsky 84] [Datskovsky 85]. A natural language interface to expert systems that would not only alleviate such limitations, but also be domain independent and easily transportable from one system to another is desirable. By studying natural language interfaces to data base systems as well as existing interfaces to expert systems, we hope to gain insights on requirements for expert system interfaces and existing natural language technology that may help us in building natural language interfaces to ex pen systems.



More About This Work

Academic Units
Computer Science
Department of Computer Science, Columbia University
Columbia University Computer Science Technical Reports, CUCS-305-87
Published Here
December 2, 2011