Sciweavers

3825 search results - page 390 / 765
» Selective Value Prediction
Sort
View
AAAI
1998
15 years 8 months ago
Bayesian Q-Learning
A central problem in learning in complex environmentsis balancing exploration of untested actions against exploitation of actions that are known to be good. The benefit of explora...
Richard Dearden, Nir Friedman, Stuart J. Russell
AAAI
1997
15 years 8 months ago
Incremental Methods for Computing Bounds in Partially Observable Markov Decision Processes
Partially observable Markov decision processes (POMDPs) allow one to model complex dynamic decision or control problems that include both action outcome uncertainty and imperfect ...
Milos Hauskrecht
ICEC
1994
37views more  ICEC 1994»
15 years 8 months ago
A Method for Utilizing Diploid/Dominance in Genetic Search
homologous alleles seemingly compete prior to manifestation in the observed phenotype, will be referred to as sub-phenotype interaction.A method is proposed for implementing diploi...
F. Greene
ENGL
2006
99views more  ENGL 2006»
15 years 6 months ago
A Patient Specific Neural Networks (MLP) for Optimization of Fuzzy Outputs in Classification of Epilepsy Risk Levels from EEG Si
In recent years Neural Networks have been widely used as pattern and statistical classifiers in bio medical engineering. Most research to date using hybrid systems (Fuzzy-Neuro) fo...
R. Sukanesh, R. Harikumar
ENGL
2007
110views more  ENGL 2007»
15 years 6 months ago
A Comparative Study of Probabilistic and Worst-case Tolerance Synthesis
— The tolerance design directly influences the functionality of products and production costs. Tolerance synthesis is a procedure that distributes assembly tolerances between com...
Jirarat Teeravaraprug