Ivan Garibay @ University of Central Florida
Abstract:
As our societies become more interconnected, understanding complex social systems has never been more important. From understanding cognitive, affective and social drivers of pandemic response behavior, to the pervasive (dis)misinformation and polarization in our social media. Generative social science is a computational approach to study complex social systems and the agent-based model is recognized as its principal scientific instrument. In generative social science, the (necessity) motto is, “if you didn’t grow it, you didn’t explain it.” In other words, given an unexplained observed macroscopic social pattern—a wealth distribution, a disease time series, a spatial segregation pattern—we seek a micro-to-macro account. Specifically, we design agents (the micro-scale) intended to generate the macro target, assessing the fit between model-generated and real-world macro structures by means of statistics. This method of agents has been successfully applied in many spheres, from epidemiology to anthropology to economics. Typically, however, agent modelers handcraft the agents, and in particular, the agents’ rules of behavior. Even when a particular model succeeds in generating the target, it is only one explanatory candidate, leaving open several questions: Is this solution unique or only one of infinitely many potential solutions? Can we find a more complete set of rules? Relatedly, can we establish a correspondence between the rules driving the computational agents and the behavioral rules at play in the real world? If we could discover a “neighborhood of” agent models whose members all generate the target, the result would seem less ad hoc, unstable and allow us to establish stronger causal connections between the model and the real word. These are the core concerns of the nascent field of Inverse Generative Social Science. In this talk, I will present Deep Agent, a modeling methodology that seeks to address these questions by leveraging artificial intelligence and large data to discover complex, robust, and causally meaningful generators to model complex social behavior. We will also discuss recent applications of these methods to social media information and misinformation diffusion and polarization, under the auspices of the Defense Advanced Research Projects Agency (DARPA) via the Computational Simulation of Online Social Behavior (SocialSim) program.
Bio:
Ivan Garibay is the director of the Complex Adaptive Systems Laboratory and the Master of Science in Data Analytics program at the University of Central Florida (UCF), where he is currently an Assistant Professor of Industrial Engineering and Management Systems. Dr. Garibay leads a team of more than 42 interdisciplinary researchers including faculty, lecturers, graduate and undergraduate students with more than $9M of extramural research funds and $1.5M/year extramural education funds. Under his leadership more than 120 student have received their master’s in data science and via his NSF-funded Innovation-Corps program, he has help more than 110 technology startups. His research expertise lies in complex systems, agent-based models, computational social science, information diffusion, information and misinformation on social media, data and network science, artificial social intelligence and machine learning. His research is currently sponsored by federal agencies and industry including NSF, DARPA, Amazon, Microsoft, Royal Bank of Canada, and the Walt Disney Corporation. He is frequently invited to present in national and international venues, including a recent congressional testimony on “Artificial Intelligence–Opportunities and Challenges Forum”, invited by US Congressman Darren Soto. He has published and presented more than 100 papers in journals and conferences.