

Concept maps have traditionally been used as a method for eliciting, representing, and storing information about domains of knowledge. In this paper, we explore an additional use of Concept Maps, to model patterns of activity in a fashion that is useful for both human understanding and computer-based deduction. Via formalization of Concept Maps the resulting information structures can be parsed automatically and used to support collaboration. The goal is to create a formalized representation of information that retains the crucial aspect of human understandability, while allowing a wide range of additional uses via adaptive system support. In this paper, we propose a process for formalizing concept maps describing a specific workflow by identifying roles, tasks, and information within the domain. This approach is used to create and store information models required by TREK, an ongoing project to support spatially-dispersed, dynamically-formed teams of workers.

