文档介绍：A KNOWLEDGE ENGINEERING APPROACH
TO NATURAL LANGUAGE UNDERSTANDING
Stuart C. Shapiro & te G. Neal
Department puter Science
State University of New York at Buffalo
Amherst, New York 14226
all knowledge, including rules, is represented as
ABSTRACT nodes in a work, an inference system
that performs reasoning according to the rules
This paper describes the results of a
stored in work, and a tracing package that
preliminary study of a Knowledge Engineering allows the user to follow the system's reasoning.
approach to Natural Language Understanding. A
A major portion of this study involves the design
computer system is being developed to handle the
and implementation of a SNePS-based system, called
acquisition, representation, and use of linguistic
the NL-system, to enable the NLU expert to enter
knowledge. puter system is rule-based and
linguistic knowledge into work in natural
utilizes a work for knowledge storage
language, to have this knowledge available to
and representation. In order to facilitate the
query and reason about, and to use this knowledge
interaction between user and system, input of
for processing text including additional NLU
linguistic knowledge puter responses are in
knowledge. These features distinguish our system
natural language. Knowledge of various types can
from other ru