Hybrid Methodology for Knowledge Elicitation and Mental Model Visualization

This Phase I project proposes to develop and demonstrate a hybrid methodology for mental model elicitation, construction, and refinement. The approach combines empirical knowledge elicitation techniques and graphical model visualization tools in a structured methodology. The knowledge elicitation techniques will consist primarily of indirect techniques, designed to access implicit and perceptually-encoded knowledge that is thought to underlie expert decision-making. Graphical displays representing the mental models will be designed based on the data obtained from battlefield domain experts. These displays will be embedded in an object-oriented interactive graphical user-interface system that will facilitate the rapid display and manipulation of these structures, including visualization of a given situation from multiple perspectives, at different points in time, and at varying levels of granularity and abstraction, to capture the types of flexible manipulation of these structures that expert decision makers are able to perform with their mental models. This work will naturally fit within a broader context of a hybrid approach to the problem of mental model determination, which would include a computational modeling component in Phase II, focusing on the refinement and validation of the inferred models through further empirical testing of model-derived hypotheses.