An analysis of linear models, linear value-function approximation, and feature selection for reinforcement learning

14 years 5 months ago

Download www.cs.duke.edu

We show that linear value-function approximation is equivalent to a form of linear model approximation. We then derive a relationship between the model-approximation error and the Bellman error, and show how this relationship can guide feature selection for model improvement and/or value-function improvement. We also show how these results give insight into the behavior of existing feature-selection algorithms.

Ronald Parr, Lihong Li, Gavin Taylor, Christopher

Real-time Traffic

Bellman Error | ICML 2008 | Linear Model Approximation | Linear Value-function Approximation | Machine Learning |

claim paper

» Feature Selection for Value Function Approximation Using Bayesian Model Selection

» Reinforcement Learning with the Use of Costly Features

» Automatic basis function construction for approximate dynamic programming and reinforcemen...

» Regularization and feature selection in leastsquares temporal difference learning

» Highaccuracy valuefunction approximation with neural networks applied to the acrobot

» DynaStyle Planning with Linear Function Approximation and Prioritized Sweeping

» Predictive State Temporal Difference Learning

» Partially supervised feature selection with regularized linear models

Post Info
More Details (n/a)

Added	17 Nov 2009
Updated	17 Nov 2009
Type	Conference
Year	2008
Where	ICML
Authors	Ronald Parr, Lihong Li, Gavin Taylor, Christopher Painter-Wakefield, Michael L. Littman

Comments (0)

Sciweavers

An analysis of linear models, linear value-function approximation, and feature selection for reinforcement learning

Bellman Error | ICML 2008 | Linear Model Approximation | Linear Value-function Approximation | Machine Learning |

Explore & Download

Productivity Tools

Document Tools

Image Tools

Sciweavers