CAKE: A Computer-Aided Knowledge Engineering Technique

Logic engineering often involves the development of modeling tools and inference mechanisms (both standard and non-standard) which are targeted for use in practical applications where expressiveness in representation must be traded off for efficiency in use. Some representative examples of such applications would be the structuring and querying of knowledge on the semantic web, or the representation and querying of epistemic states used with softbots, robots or smart devices. In these application areas, declarative representations of knowledge enhance the functionality of such systems and also provide a basis for insuring the pragmatic properties of modularity and incremental composition. In addition, the mechanisms developed should be tractable, but at the same time, expressive enough to represent such aspects as default reasoning, or approximate or incomplete representations of the environments in which the entities in question are embedded or used, be they virtual or actual. Equally important are the tools used to do the modeling. Although difficult to evaluate formally, such modeling tools should provide straightforward methods which insure the modularity and incremental composition of the knowledge structures being designed in addition to guaranteeing formal semantics and transparency of usage. The applications we are involved in require an efficient representation and query mechanism for the knowledge structures and epistemic states used by robots or softbots, in particular for applications where planning in the context of incomplete states and approximate knowledge is a necessity. We have focused on a generalization of deductive databases and query languages which involves the use of rough knowledge databases (databases where approximate relations and properties are the rule rather than the exception) and where queries can be non-monotonically contextualized to locally close only parts of the database since a closed-world assumption is not feasible. This approach provides us with a reasonably efficient query mechanism and a reasonably expressive query language for querying approximate epistemic states. In such knowledge structures, both positive and negative knowledge must be stored explicitly to ensure an open-world assumption. In the approach we are pursuing, we view a (generalized) database as a loosely coupled confederation of granular agents, where each agent is responsible for managing all or part of a relation or property. In fact, several agents may contribute locally to the definition of a relation. In addition, each relation is viewed as a partial or ap-