CIRCE: A NATURAL LANGUAGE RULE TRANSLATION SYSTEM FOR INTERSENSOR

Authors
  1. Ellis, J.R.
  2. Dent, C.A.
Corporate Authors
Defence Research Establishment Atlantic, Dartmouth NS (CAN)
Abstract
Circe is a set of functions written in LISP as an experiment in natural language processing. It was developed for knowledge acquisition by the Intersensor system, a sonar knowledge-based inference system. The main function of Circe inputs from either the keyboard or a file a rule written in English and converts it into a LISP-like data structure. This conversion takes place in two steps: the English text is first parsed, and then the relevant information is extracted from the resulting list and assembled into rules. These rules are then written to a file where they can be used directly by the inference mechanism of Intersensor without further processing. The paper briefly reviews prior research at DREA in this area, and contrasts the current system with its predecessors. Implementation problems are discussed, and suggestions for improvements are given.
Report Number
DREA-TC-85-314 — Technical Communication
Date of publication
15 Nov 1985
Number of Pages
49
DSTKIM No
86-00259
CANDIS No
105887
Format(s):
Hardcopy;Originator's fiche received by DSIS

Permanent link

Document 1 of 1

Date modified: