Modeling Human Play in Games: From Behavioral Economics to Deep Learning University of British Columbia

It is common to assume that players in a game will adopt Nash equilibrium strategies. However, experimental studies have demonstrated that Nash equilibrium is often a poor description of human players' behavior, even in unrepeated normal-form games. Nevertheless, human behavior in such settings is far from random. Drawing on data from real human play, the field of behavioral game theory has developed a variety of models that aim to capture these patterns. The current state of the art in that literature is a model called quantal cognitive hierarchy. It predicts that agents approximately best respond and explicitly model others' beliefs to a finite depth, grounded in a uniform model of nonstrategic play. We have shown that even stronger models can be built by drawing on ideas from cognitive psychology to better describe non strategic behavior. However, this whole approach requires extensive expert knowledge and careful choice of functional form. Deep learning presents an alternative, offering the promise of automatic cognitive modeling. We introduce a novel architecture that allows a single network to generalize across different input and output dimensions by using matrix units rather than scalar units, and show that its performance significantly outperforms that of the previous state of the art.