Wei Chen, Scott E. Fahlman
The ability to understand and process multiple mental contexts is an important aspect of human cognition. By "mental contexts" we mean different beliefs, states of knowledge, points of view, or suppositions, all of which may change over time. In this paper, we propose an approach for modeling and reasoning about the interactions among multiple mental contexts using the context activation scheme in Scone knowledge-base system. Our model factors the mental context representation into two separate components: (1) a dynamic mental context network (2) a set of rules which guides the activities among mental contexts and their evolution as a result of this. Our model is capable of combining newly available information and old memories stored in the context network to produce new mental states. We demonstrate the approach with a story-understanding task, in which the users feed information to the program, then ask questions about the newly updated beliefs or assumptions.
Submitted: Sep 8, 2008