SIMULATING SOCIAL INFLUENCE BASED ON REAL-WORLD GEOGRAPHIC DATA: EMERGENT NARRATIVES AND INTERACTIVE HYPOTHESIS TESTING
Abstract
Computers are increasingly being used to simulate and analyze complex social phenomena, but do not account for geographical, cultural, economic, and sociopolitical systems that influence social relationships. Social systems are complex and contextual: causes for social interaction are influenced by family of origin, economic opportunity in specific geographic regions, cultural expectations, and the means of information dissemination available. If we want to understand how social influence changes the shape of social network graphs and the course of public opinion, we cannot reason about these phenomena in isolation. We identify the need to account for real-world, localized information in social simulation, and propose to do so using open-source data sets and artificial intelligence techniques. Our objectives are to create computational models of social influence that support believable simulation and facilitate novel insights for experts through scaffolded interaction. To achieve these goals, basic research questions in simulation and interaction must be addressed. We plan to develop an interactive computational model of social influence that accounts for real-world, localized information. We will devise novel algorithms to generate virtual worlds based on real-world data, enable decision-making for virtual agents in social and geographical contexts, and model the spread of social influence. We will also develop an interaction framework to facilitate interactive hypothesis testing for social systems, supporting systems understanding about social mechanisms for novices and novel insights for experts. We expect to develop new algorithms for goal-driven decision-making, as well as better understanding of the emergent outcomes of social science theories, specifically as they predict social influence and information dissemination.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2021
- Source ID
- FA95502010355
Entities
People
- Chris Martens
Organizations
- Air Force Office of Scientific Research
- North Carolina State University
- United States Air Force