Vast resources are devoted to predicting human behavior in domains such as economics, popular culture, and national security, but the quality of such predictions is usually poor. It is tempting to conclude that this inability to make good predictions is a consequence of some fundamental lack of predictability on the part of humans. However, our recent work offers evidence that the failure of standard prediction methods does not indicate an absence of human predictability but instead reflects: 1.) misunderstandings regarding which features of human dynamics actually possess predictive power, and 2.) the fact that, until recently, it has not been possible to measure these predictive features in real world settings.
This talk introduces some of the science behind this basic observation and demonstrates its utility through three case studies. We begin by considering social groups in which individuals are influ- enced by the behavior of others; in these situations, social influence is known to decrease the ex ante predictability of the ensuing social dynamics. We show that, interestingly, these same social forces can increase the extent to which the outcome of a social process can be predicted in its very early stages. This finding is then leveraged to design prediction methods which outperform existing techniques for predicting social group dynamics.
The second case study involves analysis of the predictability of adversary behavior in the coevo- lutionary “arms races” that exist between attackers and defenders in many domains, including cyber security, counterterrorism, fraud prevention, and various markets. Our analysis reveals that conventional wisdom regarding these coevolving systems is incomplete, and provides insights which enable the development of proactive cyber defense methods that are much more effective than standard techniques. Finally, we consider the task of predicting human behavior at the level of individuals. In particular, we show that a given individual’s mobility patterns can be predicted with surprising accuracy, and conversely that knowledge of even a small portion of a person’s travel patterns permits reliable identification of that individual.
About the speaker: Rich Colbaugh received his Ph.D. in Mechanical Engineering from The Pennsylvania State University in 1986. He presently holds a joint appointment with the New Mexico Institute of Mining and Technology, where he is Chief Scientist of ICASA and a Professor in both the Mechanical Engineering and Management Departments, and Sandia National Laboratories, where he is a member of the Analytics and Cryptography Department. His research activities have focused on the modeling, analysis, and control of dynamical systems of importance in nature and society. Much of this work involves the study of very large, complex networks, including those of relevance to national security, socioeconomic systems, advanced technology, and biology.
Dr. Colbaugh spent 2001-2006 with the U.S. Intelligence Community in Washington DC advising senior leadership on counterterrorism and counterproliferation programs. Since 2007 he has concentrated his research and development efforts on social media analytics, attracting support for this program from agencies such as the Department of Defense, the Department of Homeland Security, the Department of Energy, and the National Science Foundation.