KIT Scientific Publishing

Im Shop suchen

36,00 €

inkl. 7 % USt. exkl. Versandkosten

Marc Peter Deisenroth

Efficient Reinforcement Learning using Gaussian Processes

(Karlsruhe Series on Intelligent Sensor-Actuator-Systems ; 9)

AutorDeisenroth, Marc Peter

VerlagKIT Scientific Publishing, Karlsruhe

ISBN9783866445697

UmfangIX, 205 S.

Veröffentlicht
am:
22.11.2010

Erscheinungs-
jahr
2010

VerfügbarkeitAktiv

Downloads:

Für Zitate bitte die folgende URL verwenden:
http://dx.doi.org/10.5445/KSP/1000019799

Abstract

This book examines Gaussian processes in both model-based reinforcement learning (RL) and inference in nonlinear dynamic systems.
First, we introduce PILCO, a fully Bayesian approach for efficient RL in continuous-valued state and action spaces when no expert knowledge is available. PILCO takes model uncertainties consistently into account during long-term planning to reduce model bias.
Second, we propose principled algorithms for robust filtering and smoothing in GP dynamic systems.