Predictive state representations (PSRs) have recently been proposed as an alternative to partially observable Markov decision processes (POMDPs) for representing the state of a dynamical system (Littman; Nips 2001). We present a learning algorithm that learns a PSR from observational data. Our algorithm produces a variant of PSRs called transformed predictive state representations (TPSRs). We provide an efficient principal-components-based algorithm for learning a TPSR, and show that TPSRs can perform well in comparison to Hidden Markov Models learned with Baum-Welch in a real world robot tracking task for low dimensional representations and long prediction horizons.
@INPROCEEDINGS{Rosencrantz04a,
AUTHOR = {Rosencrantz, M. and Gordon, G. and Thrun, S.},
TITLE = {Learning Low Dimensional Predictive Representations},
BOOKTITLE = {Proceedings of the Twenty-First International Conference on Machine Learning},
YEAR = {2004},
ADDRESS = {Banff, Alberta, Canada}
}