A recurrent network approach to MAP explanation

Author

Abdelbar, Ashraf M. ; Bahig, Ghada M. ; Medsker, Larry R.

Author_Institution

American Univ. in Cairo, Egypt

Volume

3

fYear

2002

fDate

6/24/1905 12:00:00 AM

Firstpage

2503

Lastpage

2507

Abstract

Bayesian belief networks (BBNs) are an increasingly popular knowledge representation for reasoning under (probabilistic) uncertainty. An important problem in BBNs is finding the best, i.e. the most probable, explanation for a given set of observations, called the evidence. In this paper, we present a recurrent neural network approach to the maximum a-posteriori (MAP) problem. We measure the performance of our approach on more than 300 pairs of belief networks and evidence sets: a combination of 23 different networks and between 10 and 21 evidence sets on each network. We find that, on average, our neural network is able to return solutions within 95% of the probability of the optimal solution

Keywords

belief networks; explanation; inference mechanisms; maximum likelihood estimation; performance evaluation; recurrent neural nets; uncertainty handling; Bayesian belief networks; MAP explanation; evidence sets; knowledge representation; maximum a-posteriori problem; most probable explanation; optimal solution; performance; probabilistic reasoning; recurrent neural network; uncertainty; Aircraft propulsion; Bayesian methods; Computer vision; Fault diagnosis; Genetic algorithms; Knowledge representation; Neural networks; Recurrent neural networks; Simulated annealing; Uncertainty;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on

Conference_Location

Honolulu, HI

ISSN

1098-7576

Print_ISBN

0-7803-7278-6

Type

conf

DOI

10.1109/IJCNN.2002.1007536

Filename

1007536