Managing Uncertainty, Importance and Differing World-Views in ASKNet Semantic Networks

ASKNet is a tool for automatically creating large scale semantic networks from natural language texts. Its semantic network formalism is robust and extensible, but as with all formalisms of this type, it is limited in the types of information it can represent. In this paper, we extend the network formalism to better represent the strength of a relation as a combination of factors such as the relation´s relevance, the reliability of the source, and our confidence that the relation is semantically correct. In addition, we extend the network to allow for post-creation adjustment of the world-view by turning off relations which were created by a particular source. This allows ASKNet to represent the information in its network as it would appear to a group or individual who either disregards, or does not have access to the information provided by a particular source.