Methods To study the ways of how real-world agents may learn to make sequential decisions, we built models