The Behavioral Modeling and Computational Social Systems group at the Georgia Tech Research Institute (GTRI) aims to explain social phenomena through understanding human behavior in computational social systems. We are interested in recovering physiological and behavioral traces of human activity that give rise to social phenomena, producing models of the behavior, and integrating behavioral models with computational social systems for predicting macro-scale behaviors.
The types of social phenomena we are interested in cut across several research areas at Georgia Tech, and the BMCSS team consists of researchers from computer science, psychology, and economics. The scope of our research touches upon several research initiatives at GTRI aimed at collecting and handling large amounts of behavioral data. The BMCSS group also collaborates with several national and international universities. The result of these collaborations is a body of work that can be used to help understand, predict, and respond to human activities so as to address fundamental national and international challenges in defense, economy, energy, health care, and education.
Specific Research Topics Include:
- Understanding patterns of Influence in Social Media using Open-Source data
- Using HPC and Large-scale Graph Analytics for Social and Cognitive Modeling
- Experimentation in Deception and Deception Detection in Social Media
- Behavioral Anomaly Detection using machine-learning methods
- Modeling of Adoption behavior at multiple scales
- Psychophysics of Influence using EEG
- Building and evaluating systems for the prediction of Emerging Technology
- Studies into the optimization of Marketing strategies
- Modeling events related to Improvised Explosive Device (IED) events
- Linking or Docking independent models for behavioral analysis
- Translating/implementing psychology and sociological theories into executable models
- Modeling behavioral influences on Terrorism and Terrorist recruitment and activities
- Representing interactions between sociological levels within models