An environment for exploring and validating declarative knowledge

Abstract

We discuss the extended LP environment of the project VEGA (knowledge validation and exploration by global analysis), a toolbox that supports the development and maintenance of declarative knowledge bases. The knowledge is represented in a declarative language that merges Horn logic with finite domains, sort hierarchies, functions, and integrity constraints. These KBs may contain problem-solving knowledge, ontologies, and cases. The VEGA environment includes tools that take into account these different kinds of knowledge to ensure the integrity of the knowledge base during its entire life-time. Besides such knowledge-validation components, VEGA provides for the interactive exploration of knowledge by inductive components, e.g. a generator of rules from facts. VEGA also allows to link foreign tools, and is thus an open architecture. The tools can be synergetically configured by visual programming. Knowledge validation can for example be performed on the output of knowledge exploration.