Learning in volatile environments with the bayes factor surprise

Surprise-based learning allows agents to rapidly adapt to nonstationary stochastic environments characterized by sudden changes. We show that exact Bayesian inference in a hierarchical model gives rise to a surprise-modulated trade-off between forgetting old observations and integrating them with th...

Full description

Bibliographic Details
Main Authors:	Brea, J. (Author), Gerstner, W. (Author), Liakoni, V. (Author), Modirshanechi, A. (Author)
Format:	Article
Language:	English
Published:	MIT Press Journals 2021
Subjects:	algorithm Algorithms animal Animals Approximate algorithms Bayes theorem Bayes Theorem Bayesian inference Bayesian networks behavior Behavior Behavioral research computer simulation Computer Simulation Economic and social effects Hierarchical model Hierarchical systems human Humans Inference engines learning Learning Learning algorithms Message passing Non-stationary environment On-line algorithms Physiological indicators physiology Reinforcement learning Reinforcement, Psychology Stochastic environment Stochastic systems Volatile environments
Online Access:	View Fulltext in Publisher

Internet

View Fulltext in Publisher

Learning in volatile environments with the bayes factor surprise

Internet

Similar Items