


Sudeep Bhatia University of Warwick, Coventry, UK Russell Golman Carnegie Mellon University
We describe the properties of a connectionist network that is able to make decisions in strategic games. We use the structure of Bidirectional Associative Memory (BAM), a minimal twolayer recurrent neural network with binary activation functions and binary connection weights. We apply BAM to finitestrategy twoplayer games, and show that network activation in the long run is restricted to the set of rationalizable strategies. The network is not guaranteed to reach a stable activation state, but any pure strategy profile that constitutes a stable state in the network must be a pure strategy Nash equilibrium.
A Recurrent Neural Network for Game Theoretic Decision Making (488 KB)