Probabilistic Inference for Fast Learning in Control

File Description SizeFormat 
springer_ewrl2008_final.pdfAccepted version260.95 kBAdobe PDFDownload
Title: Probabilistic Inference for Fast Learning in Control
Author(s): Rasmussen, CE
Deisenroth, MP
Item Type: Chapter
Abstract: We provide a novel framework for very fast model-based reinforcement learning in continuous state and action spaces. The framework requires probabilistic models that explicitly characterize their levels of confidence. Within this framework, we use flexible, non-parametric models to describe the world based on previously collected experience. We demonstrate learning on the cart-pole problem in a setting where we provide very limited prior knowledge about the task. Learning progresses rapidly, and a good policy is found after only a hand-full of iterations.
Editor(s): Girgin, S
Loth, M
Munos, R
Preux, P
Ryabko, D
Publication Date: 30-Nov-2008
URI: http://hdl.handle.net/10044/1/12223
DOI: http://dx.doi.org/10.1007/978-3-540-89722-4_18
ISBN: 978-3-540-89721-7
Publisher: Springer-Verlag
Start Page: 229
End Page: 242
Journal / Book Title: Recent Advances in Reinforcement Learning
Lecture Notes in Computer Science
Volume: 5323
Copyright Statement: © 2008 Springer-Verlag. The final publication is available at link.springer.com
Appears in Collections:Computing



Items in Spiral are protected by copyright, with all rights reserved, unless otherwise indicated.

Creative Commons