TY - GEN

T1 - An Influence Diagram framework for acting under influence by agents with unknown goals

A1 - Sønderberg-Madsen,Nicolaj

A1 - Jensen,Finn V.

AU - Sønderberg-Madsen,Nicolaj

AU - Jensen,Finn V.

PY - 2008

Y1 - 2008

N2 - <p style="text-align: left">We consider the situation where two agents try to solve each their own task in a common</p><p style="text-align: left">environment. We present a general framework for representing that kind of scenario</p><p style="text-align: left">based on Influence Diagrams (IDs). The framework is used to model the analysis depth</p><p style="text-align: left">and time horizon of the opponent agent and to determine an optimal policy under various</p><p style="text-align: left">assumptions on analysis depth of the opponent. Not surprisingly, the framework turns</p><p style="text-align: left">out to have severe complexity problems even in simple scenarios due to the size of the</p><p style="text-align: left">relevant past. We propose an algorithm based on Limited Memory Influence Diagrams</p><p style="text-align: left">(LIMIDs) in which we convert the ID into a Bayesian network and perform single policy</p><p>update. Empirical results are presented using a simple board game.</p>

AB - <p style="text-align: left">We consider the situation where two agents try to solve each their own task in a common</p><p style="text-align: left">environment. We present a general framework for representing that kind of scenario</p><p style="text-align: left">based on Influence Diagrams (IDs). The framework is used to model the analysis depth</p><p style="text-align: left">and time horizon of the opponent agent and to determine an optimal policy under various</p><p style="text-align: left">assumptions on analysis depth of the opponent. Not surprisingly, the framework turns</p><p style="text-align: left">out to have severe complexity problems even in simple scenarios due to the size of the</p><p style="text-align: left">relevant past. We propose an algorithm based on Limited Memory Influence Diagrams</p><p style="text-align: left">(LIMIDs) in which we convert the ID into a Bayesian network and perform single policy</p><p>update. Empirical results are presented using a simple board game.</p>

UR - http://pgm08.cs.aau.dk/online_proc.html

BT - Proceedings of the 4th European Workshop on Probabilistic Graphical Models

T2 - Proceedings of the 4th European Workshop on Probabilistic Graphical Models

A2 - Nielsen,Thomas D.

ED - Nielsen,Thomas D.

ER -