A Framework for Representing Knowledge
A Framework for Representing Knowledge
Toward A Model Of Children''s Story Comprehension
Toward A Model Of Children''s Story Comprehension
Hi-index | 0.00 |
This paper describes an attempt at integrating two existing methodologies to form a natural language understanding system, and at evaluating 1.PAK, an AI language developed at the University of Toronto. The design of this prototype version is based upon Fillmore's case grammar representation and Charniak's scheme for organizing routines to update a semantic network memory. Two essential components of such a system, an English parser and generation routine, have been developed separately and are not covered here, apart from describing their interfaces with the rest of the system. The system performs semantic disambiguation and deduction via demons and base routines, as in Charniak's thesis. Simple question-answering facilities are also available and planned extensions to this facility will encompass the consequent theorem methodology of PLANNER.