Adaptive Probabilistic Networks with
نویسندگان
چکیده
Probabilistic networks (also known as Bayesian belief networks) allow a compact description of complex stochastic relationships among several random variables. They are rapidly becoming the tool of choice for uncertain reasoning in artiicial intelligence. In this paper, we investigate the problem of learning probabilistic networks with known structure and hidden variables. This is an important problem, because structure is much easier to elicit from experts than numbers, and the world is rarely fully observable. We present a gradient-based algorithmand show that the gradient can be computed locally, using information that is available as a byproduct of standard probabilistic network inference algorithms. Our experimental results demonstrate that using prior knowledge about the structure, even with hidden variables, can signiicantly improve the learning rate of probabilistic networks. We extend the method to networks in which the conditional probability tables are described using a small number of parameters. Examples include noisy-OR nodes and dynamic probabilistic networks. We show how this additional structure can be exploited by our algorithm to speed up the learning even further. We also outline an extension to hybrid networks, in which some of the nodes take on values in a continuous domain.
منابع مشابه
Rule-based joint fuzzy and probabilistic networks
One of the important challenges in Graphical models is the problem of dealing with the uncertainties in the problem. Among graphical networks, fuzzy cognitive map is only capable of modeling fuzzy uncertainty and the Bayesian network is only capable of modeling probabilistic uncertainty. In many real issues, we are faced with both fuzzy and probabilistic uncertainties. In these cases, the propo...
متن کاملModeling and Control of Gas Turbine Combustor with Dynamic and Adaptive Neural Networks (TECHNICAL NOTE)
متن کامل
LPKP: location-based probabilistic key pre-distribution scheme for large-scale wireless sensor networks using graph coloring
Communication security of wireless sensor networks is achieved using cryptographic keys assigned to the nodes. Due to resource constraints in such networks, random key pre-distribution schemes are of high interest. Although in most of these schemes no location information is considered, there are scenarios that location information can be obtained by nodes after their deployment. In this paper,...
متن کاملProbabilistic Contaminant Source Identification in Water Distribution Infrastructure Systems
Large water distribution systems can be highly vulnerable to penetration of contaminant factors caused by different means including deliberate contamination injections. As contaminants quickly spread into a water distribution network, rapid characterization of the pollution source has a high measure of importance for early warning assessment and disaster management. In this paper, a methodology...
متن کاملProbabilistic Integrated Planning of Primary and Secondary Distribution Networks based on a Hybrid Heuristic and GA Approach
The integrated planning of distribution system reveals a complex and non-linear problem being integrated with integer and discontinues variables. Due to these technical and modeling complexities, many researchers tend to optimize the primary and secondary distribution networks individually which depreciates the accuracy of the results. Accordingly, the integrated planning of these networks is p...
متن کاملFuzzy completion time for alternative stochastic networks
In this paper a network comprising alternative branching nodes with probabilistic outcomes is considered. In other words, network nodes are probabilistic with exclusive-or receiver and exclusive-or emitter. First, an analytical approach is proposed to simplify the structure of network. Then, it is assumed that the duration of activities is positive trapezoidal fuzzy number (TFN). This paper com...
متن کامل