We apply the principles of communication to develop new methods for automated generation of multimedia presentations. Our focus is on natural language and information graphics explanations of quantitative and relational domains. A distinguishing feature of our approach is the principled use of intentions in planning the presentation and realizing the plan in both text and graphics. We are developing the AutoBrief system, which is a generic multimedia generation system, and using it for explanation of problems and causal relations in the domain of transportation scheduling.

We develop data exploration environments where users directly perceive and act upon information. For several years we have been collaborating with MAYA Design Group on Visage - a prototype environment, which takes a radically "information-centric" user interface approach. Visage provides techniques for locating, selecting, visualizing, manipulating, analyzing and communicating information. Visage is intended to serve as a common interface to a variety of applications providing specialized services to the user. A distinct part of Visage is VQE - a tool for visually querying complex data bases and supporting such operations as dynamic aggregation and computing derived attributes.

We want the concepts that we develop in our research to be viable and able to endure the tests of real life. Therefore, we engage in application development where we collaborate with users and demonstrate how they can perform their tasks more efficiently by using adequate visualization tools. Our first attempt at building an interface to an application system was the integration of Visage, Sage, and Autobrief into a single coherent interface of an automated reactive scheduling system (see details).