Identifying and Responding to Cooperative Actions in General-sum Normal Form Games

My primary research interest is social behavior of software agents acting in general-sum normal form games. In an environment with another agent, an agent needs to be able to identify if the opponent is hostile or cooperative by looking at the opponent’s actions, and respond appropriately. If the opponent is hostile, the agent should guard against it. Even for a self-interested agent, cooperative actions may be necessary to induce reciprocative cooperation on the part of the opponent. In some games it is easy to identify hostile or cooperative actions, but in an arbitrary general-sum game this is not easy.We introduce attitude, a method of identifying cooperative actions, and Restricted Stackelberg Response with Safety (RSRS), a solution concept for normal-form games suitable for situations where a prediction of the opponent behavior is available. Our goal is to combine attitude and RSRS to enable an agent to achieve a cooperative outcome in any general-sum game while avoiding exploitation.