Students should have a background in computer science (or a closely related field like mathematics), or psychology. Strong math skills and/or programming skills are desirable.
Computational game theory is a key tool for improving cybersecurity using artificial intelligence. It provides the capability to model interactions between multiple decision-makers, and solution concepts to analyze these models to predict likely outcomes, or evaluate courses of action. We will expand the capabilities of game theory for cybersecurity by developing and validating solution concepts based on richer concepts of human decision-making, both individually and in groups. We will conduct pilot studies to identify promising factors that are predictive of behavior, and then integrate these factors into game-theoretic models for making decisions related to cyber security. We will validate these models and the decisions they suggest using a combination of controlled experiments with human participants and analysis of data from behavioral observations of real-world users making security decisions.
The broad hypothesis of our research is that integrating findings from psychology on the influences of personality, society, and culture on decision-making into game-theoretic solution concepts will improve the ability of game theory to predict and exploit human behavior patterns in cybersecurity games.
This interdisciplinary project brings together research on psychology (particularly on dark triad personality traits) and working in computer science on computational game theory applied to cybersecurity decisions. Depending on the background of the student, they may be involved in mathematical modeling/analysis, programming, laboratory studies with human participants, and/or data analysis.