I’m Doing as Well as I Can: Modeling People as Rational Finite Automata
Professor Joseph Halpern, Professor of Mathematics and Computer Science at Cornell University. (Full Bio: https://math.cornell.edu/joseph-halpern)
Abstract: Several Nobel Prizes have been awarded for work showing that humans do not behave as maximum utility maximizers, as standard economic theory predicts. Yet they are “predictably irrational”: their deviations from rational behavior are quite systematic. Our goal is to see the extent to which we can explain and justify these deviations as the outcome of rational but resource-bounded agents doing as well as they can, given their limitations. We model people as probabilistic finite automata (a simple model of resource-bounded computation), and show that in a number of games, probabilistic finite automata doing “as well as they can” exhibit quite human-like behavior, ranging from confirmation bias to probability matching.
Moreover, this human-like behavior can actually improve performance, showing that this seemingly irrational behavior can be quite rational.
This talk covers joint work with Rafael Pass, Lior Seeman, and Lily Liu.
