Computational Kremlinology: Analyzing the Dynamics of Authoritarian Regime Elites using Networks Constructed from Images and Text
PI: Michael Gabbay, University of Washington, Applied Physics Laboratory
Co-principal Investigator: Nora Webb Williams, University of Illinois Urbana-Champaign, Department of Political Science
Years of Award: 2022-2025
Managing Service Agency: Air Force Office of Scientific Research
Shifts in elite coalitions predict changes in authoritarian regimes, including coup attempts, purges, constitutional and policy changes, as well as key appointments and demotions. Regime-affiliated media provides important public signals as to alignment and rivalry between elites and the rising or declining status of individuals and factions. In order to elucidate the factional dynamics pivotal to the politics of authoritarian regimes, this project will apply computer vision and natural language processing techniques to generate time-varying networks of elite cooperative and conflictual interactions from regime-affiliated media. Theories of authoritarian politics will be used in conjunction with network methods and complex systems models to assess individual influence, analyze regime factional structure, and capture regime dynamics of coalition formation, consolidation, and instability. We will use our data to test computational methods for the prediction of individual-level outcomes, such as promotions and sanctions, and network-level outcomes such as major factional strife, coups, constitutional changes, and contested elections.