Learning Tree Structured Potential Games

Many real phenomena, including behaviors, involve strategic interactions that can be learned from data. We focus on learning tree structured potential games where equilibria are represented by local maxima of an underlying potential function. We cast the learning problem within a max margin setting and show that the problem is NP-hard even when the strategic interactions form a tree. We develop a variant of dual decomposition to estimate the underlying game and demonstrate with synthetic and real decision/voting data that the game theoretic perspective (carving out local maxima) enables meaningful recovery.

[1]  Luis E. Ortiz,et al.  Learning the structure and parameters of large-population graphical games from behavioral data , 2012, J. Mach. Learn. Res..

[2]  Stephen P. Boyd,et al.  Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers , 2011, Found. Trends Mach. Learn..

[3]  Andrew McCallum,et al.  Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data , 2001, ICML.

[4]  Tommi S. Jaakkola,et al.  More data means less inference: A pseudo-max approach to structured learning , 2010, NIPS.

[5]  Pradeep Dubey,et al.  Strategic complements and substitutes, and potential games , 2006, Games Econ. Behav..

[6]  Elad Eban,et al.  Learning Max-Margin Tree Predictors , 2013, UAI.

[7]  Sebastian Nowozin,et al.  Structured Learning and Prediction in Computer Vision , 2011, Found. Trends Comput. Graph. Vis..

[8]  Yang Song,et al.  Optimal gateway selection in multi-domain wireless networks: a potential game perspective , 2011, MobiCom.

[9]  Alexander M. Rush,et al.  On Dual Decomposition and Linear Programming Relaxations for Natural Language Processing , 2010, EMNLP.

[10]  NowozinSebastian,et al.  Structured Learning and Prediction in Computer Vision , 2011 .

[11]  Alexander M. Rush,et al.  A Tutorial on Dual Decomposition and Lagrangian Relaxation for Inference in Natural Language Processing , 2012, J. Artif. Intell. Res..

[12]  Luis E. Ortiz,et al.  On influence, stable behavior, and the most influential individuals in networks: A game-theoretic approach , 2013, Artif. Intell..

[13]  Ben Taskar,et al.  Max-Margin Markov Networks , 2003, NIPS.

[14]  Takashi Ui,et al.  Robust Equilibria of Potential Games , 2001 .

[15]  Michael L. Littman,et al.  Graphical Models for Game Theory , 2001, UAI.

[16]  Eric P. Xing,et al.  Augmenting Dual Decomposition for MAP Inference , 2010 .

[17]  Thomas Hofmann,et al.  Large Margin Methods for Structured and Interdependent Output Variables , 2005, J. Mach. Learn. Res..

[18]  Joseph K. Bradley,et al.  Learning Tree Conditional Random Fields , 2010, ICML.

[19]  J. M. Pontiou.  On influence... , 2022, The Journal of the Oklahoma State Medical Association.

[20]  Dan Roth,et al.  Efficient Decomposed Learning for Structured Prediction , 2012, ICML.

[21]  Noah A. Smith,et al.  Dual Decomposition with Many Overlapping Components , 2011, EMNLP.

[22]  L. Shapley,et al.  Potential Games , 1994 .

[23]  Martin Hoefer,et al.  Social Context in Potential Games , 2012, WINE.