Abstract
In this chapter we will investigate the properties of observation space representations achieved by qualitative abstraction with respect to generalization and transfer learning capabilities. Section 5.1 describes the importance of structural similarity for knowledge reuse.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Baird, L.: Residual algorithms: Reinforcement learning with function approximation. In: Proceedingsof the Twelfth International Conference on Machine Learning (ICML), pp. 30–37. MorganKaufmann, San Francisco, CA (1995)
Frommberger, L.: A qualitative representation of structural spatial knowledge for robot navigationwith reinforcement learning. In: Proceedings of the ICMLWorkshop on Structural KnowledgeTransfer for Machine earning. Pittsburgh, PA, USA (2006)
Frommberger, L.: Generalization and transfer learning in noise-affected robot navigation tasks. In: Neves, J.M., Santos, M.F., Machado, J.M. (eds.) Progress in Artificial Intelligence: Proceedingsof EPIA 2007, Lecture Notes in Artificial Intelligence, vol. 4874, pp. 508–519. Springer-VerlagBerlin Heidelberg, Guimarães, Portugal (2007a)
Frommberger, L.: A generalizing spatial representation for robot navigation with reinforcementlearning. In: Proceedings of the Twentieth International Florida Artificial Intelligence ResearchSociety Conference (FLAIRS), pp. 586–591. AAAI Press, Key West, FL, USA (2007b)
Frommberger, L.: Learning to behave in space: A qualitative spatial representation for robot navigationwith reinforcement learning. International Journal on Artificial Isntelligence Tools 17(3), 465–482 (2008a)
Frommberger, L.: Representing and selecting landmarks in autonomous learning of robot navigation. In: Xiong, C., Liu, H., Huang, Y., Xiong, Y. (eds.) Intelligent Robotics and Applications:First International Conference (ICIRA 2008), Part I, Lecture Notes in Artificial Intelligence,vol. 5314,pp. 488–497. Springer Verlag Berlin Heidelberg (2008b)
Frommberger, L.: Situation dependent spatial abstraction in reinforcement learning based on structuralknowledge. In: Proceedings of the ICML/UAI/COLT Workshop on Abstraction in ReinforcementLearning. Montreal, Canada (2009)
Gordon, G.J.: Stable function approximation on dynamic programming. In: Proceedings of theTwelfth International Conference on Machine Learning (ICML), pp. 261–268. Morgan Kaufmann,San Francisco (1995)
Konidaris, G.D.: A framework for transfer in reinforcement learning. In: Proceedings of the ICML Workshop on Structural Knowledge Transfer for Machine Learning. Pittsburgh, PA, USA (2006)
Konidaris, G.D., Barto, A.G.: Autonomous shaping: Knowledge transfer in reinforcement learning. In: Proceedings of the Twenty Third International Conference on Machine Learning (ICML), pp. 489–49. Pittsburgh, PA (2006)
Roberts, F.S.: Tolerance geometry. Notre Dame Journal of Formal Logic 14(1), 68–76 (1973)
Sutton, R.S.: Generalization in reinforcement learning: Successful examples using sparse tile coding. In: Touretzky, D.S., Mozer, M.C., Hasselmo, M.E. (eds.) Advances in Neural InformationProcessing Systems: Proceedings of the 1995 Conference, vol. 8, pp. 1038–1044. MIT Press, Cambridge, MA (1996)
Timmer, S., Riedmiller, M.: Fitted Q-iteration with CMACs. In: Proceedings of the IEEE InternationalSymposium on Approximate Dynamic Programming and Reinforcement Learning, pp.1–8. Honolulu, HI (2007)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Frommberger, L. (2010). Generalization and Transfer Learning with Qualitative Spatial Abstraction. In: Qualitative Spatial Abstraction in Reinforcement Learning. Cognitive Technologies. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-16590-0_5
Download citation
DOI: https://doi.org/10.1007/978-3-642-16590-0_5
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-16589-4
Online ISBN: 978-3-642-16590-0
eBook Packages: Computer ScienceComputer Science (R0)