Kallisti seeks to develop an algorithmic theory of mind to model adversaries’ situational awareness and predict future behavior including the belief of an adversary in the state of the environment, an adversary’s risk and reward structures, and the strategies an adversary uses to attempt to project future states that achieve desired end states. The program seeks to combine algorithms with human expertise to explore, in a modeling and simulation environment, potential courses of action in national security scenarios with far greater breadth and efficiency than is currently possible. See instructions in section 1.3 to request the Controlled Unclassified Information (CUI) addendum for further program information.

Kallisti is a 30-month single-phase program organized into three sprints of 12 months, 6 months, and 12 months. DARPA anticipates selecting multiple performers. Performers must reach the goals and metrics of a Sprint to move to the next Sprint.