@InProceedings{Supelec636,
author = {Matthieu Geist and Olivier Pietquin},
title = {Statistically Linearized Least-Squares Temporal Differences},
year = {2010},
booktitle = {Proceedings of the IEEE International Conference on Ultra Modern Control systems (ICUMT 2010)},
publisher = {IEEE},
pages = {450 - 457},
month = {October},
address = {Moscow (Russia)},
url = {http://www.metz.supelec.fr/metz/personnel/geist_mat/pdfs/Supelec636.pdf},
isbn = {978-1-4244-7285-7},
doi = {10.1109/ICUMT.2010.5676598},
abstract = {A common drawback of standard reinforcement learning algorithms is their inability to scale-up to real-world problems. For this reason, a current important trend of research is (state-action) value function approximation. A prominent value function approximator is the least-squares temporal differences (LSTD) algorithm. However, for technical reasons, linearity is mandatory: the parameterization of the value function must be linear (compact nonlinear representations are not allowed) and only the Bellman evaluation operator can be considered (imposing policy-iteration-like schemes). In this paper, this restriction of LSTD is lifted thanks to a derivativefree statistical linearization approach. This way, nonlinear parameterizations and the Bellman optimality operator can be taken into account (this last point allows taking into account value-iteration-like schemes). The efficiency of the resulting algorithms are demonstrated using a linear parametrization and neural networks as well as on a Q-learning-like problem. A theoretical analysis is also provided.}
}