Sciweavers

Free Online Productivity Tools i2Speak i2Symbol i2OCR iTex2Img iWeb2Print iWeb2Shot i2Type iPdf2Split iPdf2Merge i2Bopomofo i2Arabic i2Style i2Image i2PDF iLatex2Rtf Sci2ools

10

NIPS
1994

favoriteEmaildiscussreport

178views Information Technology» more NIPS 1994»

Generalization in Reinforcement Learning: Safely Approximating the Value Function

13 years 5 months ago

Generalization in Reinforcement Learning: Safely Approximating the Value Function

Download www.ri.cmu.edu

To appear in: G. Tesauro, D. S. Touretzky and T. K. Leen, eds., Advances in Neural Information Processing Systems 7, MIT Press, Cambridge MA, 1995. A straightforward approach to the curse of dimensionality in reinforcement learning and dynamic programming is to replace the lookuptable witha generalizing function approximatorsuch as a neural net. Although this has been successful in the domain of backgammon, there is no guarantee of convergence. In this paper, we show that the combinationof dynamic programming and function approximation is not robust, and in even very benign cases, may produce an entirely wrong policy. We then introduce Grow-Support, a new algorithmwhich is safe from divergence yet can still reap the bene ts of successful generalization.

Justin A. Boyan, Andrew W. Moore

Real-time Traffic

Dynamic Programming | Neural Information Processing Systems | NIPS 1994 | NIPS 2007 | Witha Generalizing Function |

claim paper

Related Content

» Tracking value function dynamics to improve reinforcement learning with piecewise linear f...

» Parallel Reinforcement Learning with Linear Function Approximation

» Modelbased function approximation in reinforcement learning

» Kernelized value function approximation for reinforcement learning

» Gradient Descent for General Reinforcement Learning

» Efficient exploration through active learning for value function approximation in reinforc...

» Residual Algorithms Reinforcement Learning with Function Approximation

» CBR for State Value Function Approximation in Reinforcement Learning

» Value Function Approximation in Reinforcement Learning Using the Fourier Basis

Post Info
More Details (n/a)

Added	02 Nov 2010
Updated	02 Nov 2010
Type	Conference
Year	1994
Where	NIPS
Authors	Justin A. Boyan, Andrew W. Moore

Comments (0)