RECOGNITION OF LOGICAL INTERFACE STRUCTURES TO ENHANCE HUMAN-COMPUTER INTERACTION

Abstract

The objects of pattern recognition have in the past been primarily physical in nature, describable in terms of some physical units of measurement. The patterns of interest in this paper are conceptual or logical. They are to be representative of structures known to be operative in the human mind. A hybrid technique is proposed, involving the use of structural prototypes from pattern recognition combined with production systems from artificial intelligence. The purpose is to enable the computer interface to observe, detect, identify and anticipate conceptual structures. As a result, the interface system should have a more comprehensive “perspective” on what the user is doing and likely to do next. It should therefore be better prepared to provide him/her with adaptive assistance.