Abstract
Agent completely depends on trail-and-error to learn the optimal policy is the major reason to make reinforcement learning being slow and time consuming. Excepting for trail-and-error, human can also take advantage of prior learned experience to plan and accelerate subsequent learning. We propose an approach to model agent’s learning experience by Bayesian Network, which can be used to shape agent for bias exploration towards the most promising regions of state space and thereby reduces exploration and accelerate learning. The experiment results on Grid-World problem show our approach can significantly improve agent’s performance and shorten learning time. More importantly, our approach makes agent can take advantage of its learning experience to plan and accelerate learning.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction. MIT Press (1998)
Sertan, G., Faruk, P., Reda, A.: Improving reinforcement learning by using sequence trees. Machine Learning 81(3), 283–331 (2010)
Grzes, M., Kudenko, D.: Online learning of shaping rewards in reinforcement learning. Neural Networks 23(4), 541–550 (2010)
Wang, t., Daniel, L.: Bayesian Sparse Sampling for On-line Reward Optimization. In: Proceedings of the 22nd International Conference on Machine Learning, pp. 956–963 (2005)
Amizadeh, S., Ahmadabadi, M.: A Bayesian Approach to Conceptualization Using Reinforcement Learning. In: 2007 International Conference on Advanced Intelligent Mechatronics, pp. 1–7 (2007)
Doshi, F., Pineau, J.: Reinforcement Learning with Limited Reinforcement: Using Bayes Risk for Active Learning in POMDPs, pp. 256–263 (2008)
Joseph, R., Peter, S.: Online Kernel Selection for Bayesian Reinforcement Learning. In: Proceedings of the 25th International Conference on Machine Learning, pp. 816–823 (2008)
Bob, P., Craig, B.: A Bayesian Approach to Imitation in Reinforcement Learning. In: Proceedings of IJCAI 2003, Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence, pp. 712–720 (2003)
Firouzi, H., Ahmadabadi, M.N.: A Probabilistic Reinforcement-Based Approach to Conceptualization. International Journal of Intelligent Systems and Technologies 3, 48–55 (2008)
Pearl, J.: Probabilistic reasoning in intelligent systems: networks of plausible inference, pp. 79–119. Morgan Kaufmann, San Mateo (1988)
Jin, Z., Jin, J., Liu, W.: Autonomous Discovery of Subgoals Using Acyclic State Trajectories. In: Zhu, R., Zhang, Y., Liu, B., Liu, C. (eds.) ICICA 2010. LNCS, vol. 6377, pp. 49–56. Springer, Heidelberg (2010)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Jin, Z., Jin, J., Song, J. (2011). Learning Form Experience: A Bayesian Network Based Reinforcement Learning Approach. In: Liu, B., Chai, C. (eds) Information Computing and Applications. ICICA 2011. Lecture Notes in Computer Science, vol 7030. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-25255-6_52
Download citation
DOI: https://doi.org/10.1007/978-3-642-25255-6_52
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-25254-9
Online ISBN: 978-3-642-25255-6
eBook Packages: Computer ScienceComputer Science (R0)