Empirical Inference Conference Paper 2008

Policy Gradients with Parameter-based Exploration for Control

We present a model-free reinforcement learning method for partially observable Markov decision problems. Our method estimates a likelihood gradient by sampling directly in parameter space, which leads to lower variance gradient estimates than those obtained by policy gradient methods such as REINFORCE. For several complex control tasks, including robust standing with a humanoid robot, we show that our method outperforms well-known algorithms from the fields of policy gradients, finite difference methods and population based heuristics. We also provide a detailed analysis of the differences between our method and the other algorithms.

Author(s): Sehnke, F. and Osendorfer, C. and Rückstiess, T. and Graves, A. and Peters, J. and Schmidhuber, J.
Book Title: ICANN 2008
Journal: Artificial Neural Networks: ICANN 2008
Pages: 387-396
Year: 2008
Month: September
Day: 0
Editors: Kurkova-Pohlova, V. , R. Neruda, J. Koutnik
Publisher: Springer
Bibtex Type: Conference Paper (inproceedings)
Address: Berlin, Germany
DOI: 10.1007/978-3-540-87536-9_40
Event Name: 18th International Conference on Artificial Neural Networks
Event Place: Praha, Czech Republic
Digital: 0
Electronic Archiving: grant_archive
Institution: European Neural Network Society
Language: en
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik
Links:

BibTex

@inproceedings{5169,
  title = {Policy Gradients with Parameter-based Exploration for Control},
  journal = {Artificial Neural Networks: ICANN 2008},
  booktitle = {ICANN 2008},
  abstract = {We present a model-free reinforcement learning method for partially observable Markov decision problems. Our method estimates a likelihood gradient by sampling directly in parameter space, which leads to lower variance gradient estimates than those obtained by policy gradient methods such as REINFORCE. For several complex control tasks, including robust standing with a humanoid robot, we show that our method outperforms well-known algorithms from the fields of policy gradients, finite difference methods and population based heuristics. We also provide a detailed analysis of the differences between our method and the other algorithms.},
  pages = {387-396},
  editors = {Kurkova-Pohlova, V. , R. Neruda, J. Koutnik},
  publisher = {Springer},
  organization = {Max-Planck-Gesellschaft},
  institution = {European Neural Network Society},
  school = {Biologische Kybernetik},
  address = {Berlin, Germany},
  month = sep,
  year = {2008},
  slug = {5169},
  author = {Sehnke, F. and Osendorfer, C. and R{\"u}ckstiess, T. and Graves, A. and Peters, J. and Schmidhuber, J.},
  month_numeric = {9}
}