International Journal of Game Theory

, Volume 42, Issue 3, pp 695–723

Wisdom of crowds versus groupthink: learning in groups and in isolation

Article

DOI: 10.1007/s00182-012-0329-7

Cite this article as:
Mayo-Wilson, C., Zollman, K. & Danks, D. Int J Game Theory (2013) 42: 695. doi:10.1007/s00182-012-0329-7

Abstract

We evaluate the asymptotic performance of boundedly-rational strategies in multi-armed bandit problems, where performance is measured in terms of the tendency (in the limit) to play optimal actions in either (i) isolation or (ii) networks of other learners. We show that, for many strategies commonly employed in economics, psychology, and machine learning, performance in isolation and performance in networks are essentially unrelated. Our results suggest that the performance of various, common boundedly-rational strategies depends crucially upon the social context (if any) in which such strategies are to be employed.

Keywords

Bandit problems Networks Reinforcement learning Simulating annealing Epsilon greedy 

Copyright information

© Springer-Verlag 2012

Authors and Affiliations

  1. 1.Department of Philosophy, Baker Hall 135Carnegie Mellon UniversityPittsburghUSA
  2. 2.Department of Philosophy, Baker Hall 155DCarnegie Mellon UniversityPittsburghUSA
  3. 3.Department of Philosophy, Baker Hall 161ECarnegie Mellon UniversityPittsburghUSA