Neural Network Methods for Bayesian Networks
SPONSORSHIP:
NSF/IRA
BRIEF DESCRIPTION
Using Bayesian networks (AKA belief networks) to solve problems
involving causal
and probabilistic inferences has been an active research area in AI,
but existing
methods are limited because of their complexities and the difficulty in
obtaining
the needed causal knowledge. This project, sponsored by NSF, is aimed
at
overcoming these problems by developing neural network methods which
can be
directly applied to belief networks. Such integration would retain the
representational power of belief networks and at the same time take
advantage of
computational and learning capabilities of neural networks.
Specifically, we are
concentrating on two research topics here. The first one is to develop
a neural
network learning method for constructing and dynamically updating
belief networks
(both the network structures and the probability distributions) from
case data.
The second one is to adopt neural network optimization techniques to
certain
inference tasks in belief networks.
RELEVANT PAPERS
- Peng,
Y. and Jin, M.: “A Neural
Network Approach
to MAP in Belief Networks”, International Journal of Neural
Systems, 12(3, 4), 2002, 271 – 290.
- Kalra, G., Peng, Y., Guo,
M., and Augsburger, L: “A Hybrid Intelligent System
For
Formulation Of BCS Class II Drugs In Hard Gelatin Capsules”, Proceedings
of the International Conference on Neural Information Processing, Singapore, Nov. 2002.
- Peng,
Y. and Jin, M.: “A Mean
Field Approach to
MAP in Belief Networks”, in Proceedings of The
International Joint Conference on Neural Networks, Como, Italy, July, 2000.
- Peng,
Y., Jin, M., and Chen, K.: “A
Neural
Network Approach to MAPs in Belief Networks”,
in Proceedings of The International
Joint Conference on Neural Networks, Washington, DC, July 10-16, 1999.
- Peng Y,
Zhou Z, and Cho S: “Constructing Belief
Networks From Realistic Data”, International
Journal of Intelligent Systems, 14(7), 671-695, 1999.
- Peng Y
and Zhou Z: “A Neural
Network Learning
Method for Belief Networks”, International Journal of Intelligent
Systems, 11, 1996, 893-916.
- Peng Y: "Learning Probabilities for Causal networks", in
Proceedings of the International Joint Conference on Neural Networks,
Baltimore, MD, June, 1992, IV, 97-102.
- Peng Y and Jiang N: "A Learning Method for Belief Networks", The
Third International Workshop on Principle of Diagnosis, New Paltz, NY,
October, 1994.
- Peng Y: "A Neural Network Learning Method for Causal Networks",
in
Proceedings of IEEE International Conference on Systems, Man and
Cybernetics, Le Touquet, France, Oct. 1993, I, 731-736.
FOR MORE INFORMATION
Contact Yun Peng, ypeng@umbc.edu .