Algorithmic Behavior Forecasting

The success of United States'' military and Security-Stabilization-Transition-and-Reconstruction missions often hinges on the ability to understand and accurately predict the complex socio-cultural terrains of the theater of operations. The acquisition of mission-critical knowledge about the local population, conducted today by anthropological experts, is slow and prone to biases, resulting in inadequate knowledge pool, faulty recommendations, and inaccurate forecasts. To promote the rapid acquisition and effective use of the relevant socio-cultural knowledge, we propose to develop CASCADE (Computational Assessment of Socio-Cultural Action Dynamics and Events), which automates the processes of: (1) mission-critical knowledge discovery; (2) forecasting of populations’ potential responses to military-relevant events; and (3) discovery of effective means to shape populations’ responses and attitudes. CASCADE provides a unique computational framework for modeling socio-cultural terrains and dynamic behaviors and scenarios. It exploits the similarity of how behaviors are generated at the macro and micro levels to utilize the quantitative models of neural mechanisms to encode the higher-level individual and group behaviors.