Learning low dimensional predictive representations Export

@inproceedings{citeulike:3980163, abstract = {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 et al., 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. 1.}, author = {Rosencrantz, Matthew and Gordon, Geoff}, booktitle = {In ICML '04: Twenty-first international conference on Machine learning}, citeulike-article-id = {3980163}, citeulike-linkout-0 = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.10.5163}, keywords = {hierarchical}, pages = {695--702}, posted-at = {2009-01-30 01:29:24}, priority = {2}, title = {Learning low dimensional predictive representations}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.10.5163}, year = {2004} }

TY - CONF ID - citeulike:3980163 L3 - citeulike-article-id:3980163 N2 - 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 et al., 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. 1. EP - 702 TI - Learning low dimensional predictive representations T2 - In ICML ’04: Twenty-first international conference on Machine learning SP - 695 KW - hierarchical AU - Rosencrantz, Matthew AU - Gordon, Geoff PY - 2004/// UR - http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.10.5163 ER -