This is a project to advance synthetic character technology as a novel platform for assessing adolescents' decision making and social competency skills. Virtual vignettes will present the adolescent with full-body animated, conversant characters with whom the adolescent interacts via natural language and who exhibit emotional, gestural, cognitive, and social intelligence. Because inputs and responses to the synthetic characters are natural, the vignettes assess adolescents' situation-specific behavior rather than test their understanding of how they think they should behave.

This research builds on a recently completed study that provided preliminary evidence that virtual vignettes can yield performance measures with construct and criterion validity. It will investigate how to improve the fidelity of the synthetic characters in developing more appropriate language, gesture, cognitive, and social models to drive their behavior. These activities will require novel approaches to linguistic, emotional, gestural, cognitive, and social modeling. In-depth experimental analyses (usability and accessibility studies, acceptance, and performance measuring) will compare adolescents who exhibit conduct disorder with matched normal adolescents using the application, focusing on how the different groups attend differently to virtual stimuli, to determine what causes differences in behavior such as acquiescing in risky behavior. The research will also explore how well the technology might extend towards assessment of behavior of different participant groups.

This research addresses the need to empower a traditionally underserved group with new combinations of computer interfaces, including virtual reality, natural language, and emotive synthetic characters. The work will advance the knowledge and understanding of both synthetic character technology and the needs of at-risk children, and particularly how researchers should go about developing collaborative, adaptive systems that address adolescents' needs. Ultimately, tools based on synthetic character technology could assess poor social skills and offer educational experiences within the classroom, clinical settings, and juvenile correctional institutions. In the long term, the techniques may be adapted as early interventional measures for younger children who may have cognitive dysfunctions but who have not yet developed inappropriate behaviors.

Agency
National Science Foundation (NSF)
Institute
Division of Information and Intelligent Systems (IIS)
Application #
0534211
Program Officer
William Bainbridge
Project Start
Project End
Budget Start
2006-06-15
Budget End
2009-05-31
Support Year
Fiscal Year
2005
Total Cost
$387,962
Indirect Cost
Name
Research Triangle Institute
Department
Type
DUNS #
City
Research Triangle Park
State
NC
Country
United States
Zip Code
27709