The Support of Higher-level Cognition in the Context of Ill-structured Process Knowledge

How can knowledge acquisition, modeling, problem solving, and discovery be facilitated in domains where systems involve complex, heterogeneous, and multi-scale processes? This paper presents the conceptual developments of a research program that aims to support higher-level cognition in such demanding contexts. These process domains present a major challenge for human cognition because the related knowledge is ill-structured and spans different scientific areas that employ a wide spectrum of paradigms and models. Further, diverse notations are required to encode the knowledge. Our approach describes a problem space for performing knowledge acquisition in such domains and is set in a wider context of modeling and discovery. We have developed a framework of structured representations, methods, and heuristics, which are utilized to elicit and encode ill-structured process knowledge in a largely diagrammatic knowledge-based system, and to provide human-level interfaces for an interactive software program, designed to aid humans in the specification, inspection, and exploration of ill-structured process knowledge. Specially designed diagrams are fundamental to the approach because they provide a neutral notation to integrate many aspects of heterogeneous knowledge, whilst exploiting cognitive and representational properties that provide advantages for human understanding and reasoning. Evidence for the utility of the approach was found as a result of an initial knowledge acquisition case study in a representative industrial domain that involves combinations of physical, chemical, and biological processes.