|
|
|
|
|
• By Date • By Publication Type • By Research Category • By Author (Currently at CMU) • By Author (Past and Current CORAL Contributors) •
S. Chernova and M. Veloso. Tree-Based Policy Learning in Continuous Domains through Teaching by Demonstration. In Modeling Others from Observations: Papers from the AAAI Workshop, American Association for Artificial Intelligence, Menlo Park, California, July 2006.
This paper addresses the problem of reinforcement learning in continuous domains through teaching by demonstration. Our approach is based on the Continuous U-Tree algorithm, which generates a tree-based discretization of a continuous state space while applying general reinforcement learning techniques. We introduce a method for generating a preliminary state discretization and policy from expert demonstration in the form of a decision tree. This discretization is used to bootstrap the Continuous U-Tree algorithm and guide the autonomous learning process. In our experiments, we show how a small number of demonstration trials provided by an expert can significantly reduce the number of trials required to learn an optimal policy, resulting in a significant improvement in both learning efficiency and state space size.
@inproceedings{Chernova2006moo,
author = {S. Chernova and M. Veloso},
booktitle={Modeling Others from Observations: Papers from the AAAI Workshop},
title ={Tree-Based Policy Learning in Continuous Domains through Teaching by Demonstration},
editor = {Gal Kaminka and David Pynadath and Christopher Geib},
address = {Menlo Park, California},
publisher = {American Association for Artificial Intelligence},
year = {2006},
abstract={
This paper addresses the problem of reinforcement learning in continuous domains through teaching by demonstration. Our approach is based on
the Continuous U-Tree algorithm, which generates a tree-based discretization of a continuous state space while applying general reinforcement
learning techniques. We introduce a method for generating a preliminary state discretization and policy from expert demonstration in the form
of a decision tree. This discretization is used to bootstrap the Continuous U-Tree algorithm and guide the autonomous learning process. In
our experiments, we show how a small number of demonstration trials provided by an expert can significantly reduce the number of trials
required to learn an optimal policy, resulting in a significant improvement in both learning efficiency and state space size.},
bib2html_pubtype = {Workshop},
bib2html_rescat = {Robot Learning, Learning from Demonstration}
bib2html_dl_pdf = {http://www.cs.cmu.edu/~soniac/files/ChernovaVelosoMOO_06.pdf},
}
Generated by bib2html.pl (written by Patrick Riley ) on Tue Oct 09, 2007 00:00:13
|
|
 | |||
| [home] [news] [people] [robosoccer] [projects] [multimedia] [download] [publications] [contact us] | |||