The volume of information available is making it impractical to rely completely on humans to touch, assess and classify every bit of information, especially when there may be multiple overlapping security classification guidelines that are applicable. Automation is needed on two fronts: (1) to design security classification guides that are clear, formal, explicit and mutually consistent and non-redundant (2) to apply security classification guides with the smallest amount of feasible human intervention. We will use an existing language to logic translation system (called the Controlled English to Logic Translation system) to convert security classification guides into formal logic, utilizing new and existing definitions from the Suggested Upper Merged Ontology (SUMO). We will ensure that the translated rules are clear and consistent with automated theorem proving conducted in our Sigma Knowledge Engineering Environment.
Benefit: The resulting system will enable security classification guides to be turned into computer-executable specifications. Security guides will be able to be tested automatically for internal consistency and consistency with other guides.
Keywords: Ontology, Automated Reasoning, Natural Language Understanding, Logic, Knowledge Representation
