Model Minimization For Online Predictability

For humans in a teaming scenario, context switching between reasoning about a teammate's behavior and thinking about thier own task can slow us down, especially if the cognitive cost of predicting the teammate's actions is high. So if we can make the prediction of a robot-teammate's actions quicker, then the human can be more productive. In this paper we present an approach to constrain the actions of a robot so as to increase predictability (specifically online predictability) while keeping the plan costs of the robot within acceptable limits. Existing works on human-robot interaction do not consider the computational cost for predictability, which we consider in our approach. We approach this problem from the perspective of directed graph minimization, and we connect the concept of predictability to the out-degree of vertices. We present an algorithm to minimize graphs for predictability, and contrast this with minimization for legibility (goal inference) and optimality.