Visualization has become an indispensable tool in many areas of science and engineering. Advanced visualization techniques allow scientists to view and explore their computational results, but truly effective systems allow the discovery of unexpected and often subtle aspects of the data. Such discoveries can only be made by those intimately familiar with the generation of the data. Adding to the visualization system the capability to learn from the domain experts in complex data analysis tasks can facilitate similar tasks and increase the utilization and efficacy of the system. This research aims to create such a new visualization technology that is anticipated to significantly lower the cost of visualization.
The investigators study how to integrate ?intelligence? into visualization systems to automatically handling simple or repetitive tasks, and to effectively assist users in performing complex tasks involving large, high-dimensional data. Only high-level, goal-oriented decisions need to be made by the user, making cutting-edge visualization technology directly accessible to a wide range of application scientists. One research task is to select suitable machine learning methods for representative visualization tasks. Two demanding visualization applications, turbulent flow analysis and social data analysis, are used in this study. The other critical task is to consult domain experts for understanding the visualization task requirements and visual language, followed by the design of an appropriate user interface for each visualization task. The research results can therefore help realize truly coherent and usable visualization systems and broaden the base of potential users of visualization technology.
