The CESAR lab works on the computational modeling of human behavior, both as a basic research method in the study of human behavior as well as the use of these computational models in a range of education and analysis applications. Current research includes modeling human decision-making, social interaction and nonverbal behavior at the individual, dyadic and group level. Specifically, at the individual level, we are using Bayesian approaches to model the relation between the perception of nonverbal expressions, inferences of underlying emotional states and theory of mind reasoning about others. We are also exploring data-driven approaches to making inferences about personality from objective, behavioral measures, using smart phone technology. At the dyadic level, we are employing experimental techniques and data driven approaches to study and model human social interaction in negotiation contexts as well as human-agent social interaction in experimental games and virtual reality. At the multi-agent level, we are using data driven approaches to develop models of human behavior in large scale social technical systems under stress from disasters and disruptions.
The lab has a special interest in the application of these models to the design of virtual humans and social simulations for training. Specifically, the current research is being applied to craft health interventions, simulate supply chains, study the response to large scale disasters and crowd agent-based interactive narratives for social skills training.