Open Collections

Deep learning for predicting human strategic behavior

University of British Columbia

Predicting the behavior of human participants in strategic settings is an important problem for applications that rely on game theoretic reasoning to design mechanisms or allocate resources. Most existing work either assumes that participants are perfectly rational, or attempts to directly model each participant's cognitive processes based on insights from cognitive psychology and experimental economics. In this work, we present an alternative, a deep learning-based approach that automatically performs cognitive modeling without relying on such expert knowledge. We introduce a novel architecture that allows a single network to generalize across normal form games with varying numbers of actions. We show that the architecture generalists the most successful existing models and that its performance significantly improves upon that of the previous state of the art, which relies on expert-constructed features.

Text

2016-10-25

Attribution-ShareAlike 4.0 International

http://hdl.handle.net/2429/59559

Computer Science

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-sa/4.0/