neural networks research group
areas
people
projects
demos
publications
software/data
An Empirical Comparison of Abstraction in Models of Markov Decision Processes (2009)
Todd Hester
and
Peter Stone
Reinforcement learning studies the problem of solving sequential decision making problems. Model-based methods learn an effective policy in few actions by learning a model of the domain and simulating experience in their models. Typical model-based methods must visit each state at least once, which can be infeasible in large domains. To overcome this problem, the model learning algorithm needs to generalize knowledge to unseen states and provide information about the states in which it needs more experience. In this paper, we use existing supervised learning techniques to learn the model of the domain. We empirically compare their effectiveness at generalizing knowledge across states on three different domains. Our results indicate that tree-based models perform the best after training on a small number of transitions, while support vector machines perform the best after a large number of transitions.
View:
PDF
,
PS
,
HTML
Citation:
In
Proceedings of the ICML/UAI/COLT Workshop on Abstraction in Reinforcement Learning
, June 2009.
Bibtex:
@InProceedings{ICMLARL09-hester, title={An Empirical Comparison of Abstraction in Models of Markov Decision Processes}, author={Todd Hester and Peter Stone}, booktitle={Proceedings of the ICML/UAI/COLT Workshop on Abstraction in Reinforcement Learning}, month={June}, url="http://nn.cs.utexas.edu/?ICMLARL09-hester", year={2009} }
People
Todd Hester
todd [at] cs utexas edu
Peter Stone
pstone [at] cs utexas edu
Projects
TEXPLORE: Real-Time Sample Efficient Reinforcement Learning
Since 2009
Demos
TEXPLORE: Real-Time Sample Efficient Reinforcement Learning
Todd Hester
2012
Areas of Interest
Reinforcement Learning