Knowledge packets and knowledge packet structures

One of the most challenging tasks in knowledge representation for text-understanding systems is the development of large-scale knowledge bases containing semantic background knowledge. Many current knowledge-based systems restrict their applicational scope to narrow world domains, that is, they are equipped to handle domain-specific tasks by using domain-specific facts and rules. However, for the domain of text understanding this restriction appears to be unacceptable since the understanding of naturallanguage texts draws on a large background of diverse world knowledge and belief. Even if a knowledgebased system could be provided with a large body of world knowledge sufficiently rich to cover a broader domain of discourse, the problem of using this knowledge to build and utilize representations for natural language inputs might prove a serious one with respect to computational resources.