Policy gradient methods are a type of reinforcement learning techniques that rely upon optimizing parametrized policies with respect to the expected return (long-term cumulative reward) by gradient descent. They do not suffer from many of the problems that have been marring traditional reinforcement learning approaches such as the lack of guarantees of a value function, the intractability problem resulting from uncertain state information and the complexity arising from continuous states & actions.
| Author(s): | Peters, J. |
| Journal: | Scholarpedia |
| Volume: | 5 |
| Number (issue): | 11 |
| Pages: | 3698 |
| Year: | 2010 |
| Month: | November |
| Day: | 0 |
| Bibtex Type: | Article (article) |
| DOI: | 10.4249/scholarpedia.3698 |
| Digital: | 0 |
| Electronic Archiving: | grant_archive |
| Language: | en |
| Organization: | Max-Planck-Gesellschaft |
| School: | Biologische Kybernetik |
| Links: | |
BibTex
@article{6940,
title = {Policy gradient methods},
journal = {Scholarpedia},
abstract = {Policy gradient methods are a type of reinforcement learning techniques that rely upon optimizing parametrized policies with respect to the expected return (long-term cumulative reward) by gradient descent. They do not suffer from many of the problems that have been marring traditional reinforcement learning approaches such as the lack of guarantees of a value function, the intractability problem resulting from uncertain state information and the complexity arising from continuous states & actions.},
volume = {5},
number = {11},
pages = {3698},
organization = {Max-Planck-Gesellschaft},
school = {Biologische Kybernetik},
month = nov,
year = {2010},
slug = {6940},
author = {Peters, J.},
month_numeric = {11}
}