For the past dozen years, a team at RTI has developed behavioral software and applications that enable synthetic characters to act and behave realistically in controlled learning contexts, for interaction skills training and assessment. Applications include training civilian police officers in how to handle encounters with mentally disturbed individuals, assessing medical practitioners in history taking for both asthmatic and pediatric patients, monitoring the transmission and verification of Intelligence information, and training telephone and field interview staff in obtaining respondent participation. Much of the research has involved carefully-designed dialog models to enable natural flow and the eliciting or provision of necessary information between conversational partners. This presentation will address dialog design issues including coverage of content, consideration of social roles, biased reasoning, sensitive topics, and lexical and syntactic constraints.

Rob Hubal, a senior researcher at RTI International, is interested (1) in research focusing on development, presentation, and evaluation of learning materials and (2) identifying approaches to improve learning and training effectiveness. He is interested in applying research results to both everyday and specialized domains. Dr. Hubal holds an M.S. degree in Computer Science from North Carolina State University and a Ph.D. in Cognitive Psychology from Duke University.