Proceedings of Human Social Culture Behavior (HSCB) Focus 2011, Chantilly, VA (February 2011)
Human socio-cultural behavior (HSCB) models are proliferating as a means to examine and reason about human behavior across many environments and for many different purposes. Prior work has identified issues surrounding operator trust in those models, including guidelines for the design of appropriate trust guided by prior research on trust in automated systems. HSCB models may be engender inappropriate levels of trust because they are not typically “one-size-fits-all;” different theories and modeling methods are likely to be required across different operational domains. The explicit visual representation of trust-related factors may help operational users determine what model or set of HSCB models are appropriate for a given use case, how to parameterize or situate those models, how to understand the relevant internal states of the model, and how to interpret model results. In this paper, we present a list of potential trust-related factors, their relationship to cognitive tasks, and techniques to represent those factors visually to users. Trust-relevant factors include meta-information about a model’s previous performance and accuracy under different conditions, and traces of a model’s operation to provide insight and observability about the model. We explore leading visualization techniques in spatio-temporal, network, and causal contexts to represent trust-relevant factors. We also provide a set of case studies that demonstrate how operational tasks, trust issues, and effective representations can be combined to enhance operational performance and foster appropriate trust.
For More Information
(Please include your name, address, organization, and the paper reference. Requests without this information will not be honored.)