This project will develop a computer system which will solve problems creatively and will be based on the extension, abstraction and generalization of human cognitive processes. A model of cognitive processes exists as a computer program for simple knowledge-based problem solving using a highly parallel architecture with components for processing goals, plans, memory and perception. In order to extend this model it will be necessary to extend its various components to incorporate new types of knowledge that are to be used creatively and to introduce new mechanisms for conceptual representations and their manipulation. The new method of representing knowledge allows knowledge to be flexibly abstracted or constrained as required to discover new problem solutions. The highly parallel nature of the existing model will allow many different kinds of knowledge to be activated concurrently thereby allowing a large set of potential solutions to be considered, but it will also allow a large set of constraints to be applied concurrently to find the best solutions.
The intellectual merit of the project will be in advancing our understanding of the important area of creativity by showing how different knowledge-based processes work together to creatively solve problems. Its broader impact will be in showing how to develop systems which support and enhance human creativity in the design of commercial products, in artistic and cultural activities, in understanding the development of creativity in our children, in educating our students to be more creative, and to have marketable creative skills.
