name: file RA2_CWL_lab-stmtn.doc date last saved: 2005/10/07 9:09
Page 1 of 6
Research Area #2: Visual representations and interaction techniques
The successful development of innovative visual analytic environments must incorporate hard won lessons from cognitive science and human-computer interaction that human cognition is embodied cognition and that interface design has a large role to play in increasing productivity and reducing human error. We propose to address the issues presented by visual representations and interaction techniques through a combination of basic research, applied research, and the development of research results into tools for cognitive engineering visual task environments. In doing so, our proposal will directly addresses many of the recommendations listed in Chapter 3 of Illuminating the Path. Our approach is based in cognitive science theory, directed by human factors sensibilities, and honed by ARDA-sponsored research with intelligence analysts. Our research tools will feedback into our basic and applied research efforts. These tools will be designed to meet three sets of needs. First is the use of families of near-optimal performance models (simulated cyborgs or simBorgs) that combine cognitively plausible computational models of interactive behavior with artificial intelligence based reasoning components. Our goal is the creation of intelligent agents, simBorgs, that perform usability testing on various combinations of tasks and interfaces with human time constants. Second is a program of experimental design and empirical data collection and analysis that combines traditional performance and response time measures with an analysis of the microstructure of cognition, perception, and action at the 16.666 millisecond level of analysis. Third is the building and dissemination of cognitively engineered software systems that permit our data collection tools and simBorgs to easily instrument the complex visualization software designed by other groups. Such use entails both empirical data collection, exploratory data analysis, and the use of simBorgs to perform the same tasks using the same software as human users. For an immediate start, we will begin by enhancing the CogWorks Visualization-Interaction Architecture (VIA) that was successfully deployed to the University of Maryland this summer where it was used to instrument TreePlus (developed by a graduate student of Bederson and Plaisant at the HCIL). VIA is a cognitive engineering tool that automates all empirical data collection, facilitates the use of visualization software by computational cognitive models, and provides facilities for exploratory analysis of high density data as well as the analysis of dynami...
RA2_CWL_lab-stmtn