COGNET and BATON: an integrated approach to embedding user models in complex systems

The authors discuss conceptual and computational aspects of the COGNET (cognitive network of tasks)/BATON (backboard architecture for task-oriented networks) architecture for embedding user models in an operational system or simulation. Technology to capture and embed cognitive models of the system user in an active HCI (human-computer interaction) is key for intelligent human-computer integration. There are two serious problems that this enabling technology must overcome: how to built a model of the human user that is sufficiently powerful and accurate that it can act as an embedded model subsystem, given the early and imprecise state of the art of cognitive modeling; and how to implement such a model computationally, once it is created, in the robust and efficient form needed for real-time interaction. The application of the COGNET model to built a user model and the implementation of the model using BATON are considered. An example application for active HCI from the ASW (air antisubmarine warfare) problem domain is discussed