Concept-Level Rules for Capturing Domain Knowledge

Formal ontology and rule-based approaches founded on semantic technologies have been gaining traction in capturing domain knowledge in numerous fields. This typically requires several iterations between a Subject Matter Expert (SME) and a modeling expert to capture domain knowledge; resulting in a time-consuming process. Recent advances in formal Controlled Natural Languages (CNL) improve the process as the SME can more easily evaluate and verify the captured domain knowledge. Such executable domain rules are usually represented in some encoding / representation language and typically use variables. This use of variables does not correspond to how we naturally think about a domain and hence the difficulty that SMEs encounter in directly authoring domain rules. Our goal is to enable the SMEs to directly author the rules in a formal yet English-like language. For this purpose, we introduce the notion of Concept Rules (Crules) where the domain rules are captured in terms of concepts and relationships without the need to use variables. The absence of variables makes the captured rules much more English-like and natural and correspond more closely with how we communicate in our day-to-day life. Use of Crules will make it easier for an SME to directly author domain rules. The Crules captured are automatically converted to a target language to leverage existing execution platforms.