Predicting Human Interactive Learning by Regret-Driven Neural Networks

by: Davide Marchiori, Massimo Warglien

Science, Vol. 319, No. 5866. (22 February 2008), pp. 1111-1113.

View FullText article

DOI, HighWire, Pubmed, Hubmed

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

mdimmic, brian, Koh, shigepong, zwang, shivakmr, VisionLab (group)

Find related articles with these CiteULike tags

decision-making, economics, game, interaction, learning, learningingames, network, neural, neural-models, prediction, psychology, regret, science, shiva

Abstract

Much of human learning in a social context has an interactive nature: What an individual learns is affected by what other individuals are learning at the same time. Games represent a widely accepted paradigm for representing interactive decision-making. We explored the potential value of neural networks for modeling and predicting human interactive learning in repeated games. We found that even very simple learning networks, driven by regret-based feedback, accurately predict observed human behavior in different experiments on 21 games with unique equilibria in mixed strategies. Introducing regret in the feedback dramatically improved the performance of the neural network. We show that regret-based models provide better predictions of learning than established economic models. 10.1126/science.1151185

@article{citeulike:2409463, abstract = {Much of human learning in a social context has an interactive nature: What an individual learns is affected by what other individuals are learning at the same time. Games represent a widely accepted paradigm for representing interactive decision-making. We explored the potential value of neural networks for modeling and predicting human interactive learning in repeated games. We found that even very simple learning networks, driven by regret-based feedback, accurately predict observed human behavior in different experiments on 21 games with unique equilibria in mixed strategies. Introducing regret in the feedback dramatically improved the performance of the neural network. We show that regret-based models provide better predictions of learning than established economic models. 10.1126/science.1151185}, author = {Marchiori, Davide and Warglien, Massimo }, citeulike-article-id = {2409463}, doi = {10.1126/science.1151185}, journal = {Science}, keywords = {interaction, network, neural, prediction}, month = {February}, number = {5866}, pages = {1111--1113}, posted-at = {2008-04-12 13:33:57}, priority = {4}, title = {Predicting Human Interactive Learning by Regret-Driven Neural Networks}, url = {http://dx.doi.org/10.1126/science.1151185}, volume = {319}, year = {2008} }

RIS record

TY - JOUR ID - citeulike:2409463 L3 - citeulike-article-id:2409463 TI - Predicting Human Interactive Learning by Regret-Driven Neural Networks JF - Science VL - 319 IS - 5866 SP - 1111 EP - 1113 N2 - Much of human learning in a social context has an interactive nature: What an individual learns is affected by what other individuals are learning at the same time. Games represent a widely accepted paradigm for representing interactive decision-making. We explored the potential value of neural networks for modeling and predicting human interactive learning in repeated games. We found that even very simple learning networks, driven by regret-based feedback, accurately predict observed human behavior in different experiments on 21 games with unique equilibria in mixed strategies. Introducing regret in the feedback dramatically improved the performance of the neural network. We show that regret-based models provide better predictions of learning than established economic models. 10.1126/science.1151185 KW - interaction KW - network KW - neural KW - prediction AU - Marchiori, Davide AU - Warglien, Massimo PY - 2008/02/22/ UR - http://dx.doi.org/10.1126/science.1151185 ER -

RIS BibTeX