CCN 2019: Using Inverse Reinforcement Learning to Predict Goal directed Shifts of Attention

Gregory Zelinsky, Stony Brook University, United States Understanding how goal states control behavior is a question intersecting attention, action, and recognition, and one that is ripe for interrogation by new methods from machine learning. This study uses inversereinforcement learning (IRL) to learn the reward function and policy underlying the simplest of goaldirected actionsshifts of gazein the service of the simplest of goalsfinding a desired target category. Training this IRL model of categoric