Computational Models of Human Behavior

Computational models (mathematical representation) of people's behaviors enable User Interfaces (UI) that can automatically reason about common user behaviors, understand and infer user goals, predict future user actions, and even coach users when using the interfaces. However, today's models often make specialized prediction and classification without regard for underlying human behavior which leads to erroneous representations of the behavior (e.g., models with missing parameters, models built on spurious correlations in behavior traces data) and in turn wrong prediction and misclassification. This is a common pitfall when modeling behavior from traces stored in behavior logs containing large amount of unlabeled, heterogeneous data, which cannot be easily understood using traditional data exploration approaches. To address those challenges, I apply a methodology to create models grounded in theoretical knowledge about human behavior that not only predict behaviors, but that can also explain and describe those behaviors to minimize such errors. I analyze people's behavior to formalize it in a mathematical explanation, which I implement as a computational model. I validate the model by ensuring that it fits empirical behavior data. My models synthesize quantitative data from behavior logs that can be collected using today's sensing technologies into a format that can easily be visualized and analyzed. I illustrate the applications of computational models of human behavior in domains of Computational Interaction, Behavior Data Analytics, and Human-Data Supported Interfaces. I conclude with a vision for modeling the human accurately across domains as a theoretical foundation for work in Human-Computer Interaction in which computational models provide a quantitative method to study, describe, and understand complex human behaviors and support design, optimization, and evaluation of user interfaces.
Nikola Banovic is a Ph.D. candidate at the Human-Computer Interaction Institute at Carnegie Mellon University. His research broadly focuses on understanding and modeling human behavior to support innovative information technology that will change how we study and design interactive user experiences to have a positive impact on society through improving the quality of people's lives. Nikola received his Honours Bachelor of Science and Master of Science degrees from the University of Toronto. Nikola is an NSERC Post-graduate Fellow and a Yahoo! Fellow, and has received four paper awards for his work on modeling and understanding people's behaviors.