Multi-Agent Framework
Each simulated entity (individual, group, or organization) operates with private beliefs, goals, emotional states, and decision-making heuristics. Together, agents produce realistic group dynamics, social influence, and emergent collective behavior across teams, organizations, and larger systems.
Hybrid Archetecture
We combine model-based and attention-based generative approaches. Model-based methods ground simulations in evidence-based research, align them with doctrine, and provide users with control and transparency, while generative methods expand fidelity, reduce authoring overhead, and support more natural interaction.
Theory-driven Computational Model
We build computational models from social science, affective science, cognitive theories of persuasion, and behavioral economics to represent how people think, feel, interact, and influence one another. Grounded in decision theory, stochastic planning, and sequential decision making, these agents respond to scenarios with the cognitive, social, and emotional nuance of real individuals and groups.
Validated Basic and Applied Research
Our core technology has supported simulation-based decision aids and training across domains, including negotiation, cyber security, search and rescue, disaster response, influence campaigns, and urban stabilization. Backed by agencies including DARPA, ARL, SOCOM, ONR, NSF, and NIMH, these systems enabled experimentation of tactics and strategies before applying them in the real world.