J. H. Kim and J. Peal. A computational model for causal and diagnostic reasoning in inference systems. In Proceedings of the 8th International Joint Conference on Artificial Intelligence, pages 190--193. 1983.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....document represented as a Direct Acyclic Graph. This document is composed of an introduction and two sections. The fist section has two paragraphs and the second one. Each part of the document is represented by a node with a label and a textual information 4. 1 Belief networks Belief networks [9] are stochastic models for computing the joint probability dis tribution over a set of random variable. They are DAGs whose nodes are the random variables and edges correspond to probabilistic dependence relations between 2 variables. The structure of the DAG reflects conditional independence ....

Jin H. Kim and Judea Pearl. A Computational Model for Causal and Diagnostic Reasoning in Inference Systems. In Alan Bundy, editor, Proceedings of the 8th International Joint Conference on Artificial Intelligence, Karlsruhe, Germany, August 1983. William Kaufmann.

....for more detail. 3 BAYESIAN NETWORK INFERENCE ALGORITHMS REVIEW In this section, we will briefly review exact and approximate Bayesian Networks inference algorithms in general. 3. 1 Exact Inference In early 1980s, Pearl published efficient message propagation inference algorithm for polytrees [KP83, Pe86a, Pe86b]. The algorithm is exact and has Cloudy Rain Sprinkler WetGrass P(C) 50 CP(R) CP(S) S R P(S) TT .10 TF .50 FT .90 FF .00 polynomial complexity in the number of nodes, but works only for polytrees. Pearl also presented an exact inference algorithm for multiply connected networks ....

J. H. Kim and J. Pearl. A computational model for causal and diagnostic reasoning in inference engines. In Proceedings of the 8th International Joint Conference on Artificial Intelligence, Karlsruhe, West Germany, pages 190--193. International Joint Conference on Artificial Intelligence, August 1983.

....reasoning with the ID is also distributed, since each agent reasons with its part of the ID. We use a variation of the local conditioning (LC) algorithm [13] for infering in a distributed manner. LC is an extension of the classical algorithm for evidence propagation for single connected networks [22]. The basis of LC is the process of building an associated tree to the ID. The tree is build by doing a depth first search in the ID as a way to detect and break the loops in the network. Then, in the resulting tree, a simple variation of the distributed LC algorithm [13] for evidence propagation ....

J. H. Kim and J. Pearl. A computational model for causal and diagnostic reasoning in inference engines. In International Joint Conference on Artificial Intelligence, pages 190--193. Morgan Kaufmann, 1983.

....tradeoffs among these attributes for an individual patient, as well as individual preferences about risk and time. Much of the research on representation and inference with these graphical representations has focused on specializations of influence diagrams that contain only chance nodes [126, 82, 97, 21, 116, 89]. These express probabilistic relationships among states of the world exclusively, without explicit consideration of decisions and values. Several different terms are used for these representations, including causal networks, Bayesian nets, and belief networks [114] We use belief networks, as ....

....to avoid having to calculate explicitly the full joint probability distribution. A variety of methods have been developed, each focusing on particular families of belief network topology. Kim and Pearl have developed a distributed algorithm for solving singly connected networks, or polytrees [89]. The algorithm is linear in the number of variables in the network. In this scheme, each node in the network obtains messages from each of its parent and child nodes, representing all the evidence available from alternative portions of the network. The single connectedness guarantees that the ....

J.H. Kim and J. Pearl. A computational model for causal and diagnostic reasoning in inference engines. In Proceedings of the 8th International Joint Conference on Artificial Intelligence, Karlsruhe, West Germany, pages 190--193. International Joint Conference on Artificial Intelligence, August 1983.

....applies. The main purpose of this paper is to document a connection between constraint reasoning and probabilistic reasoning. We show that pAC algorithm is a generalization of the basic arc consistency algorithm AC 3 [10] and is also a specialization of the belief propagation algorithm [8] for singly connected Bayesian networks. However, since the value of our method must be established empirically, we will briefly describe some of our empirical results in Section 6. We feel our results are positive: we can report a dramatic decrease in search costs, i.e. number of backtracks, ....

....belief propagation algorithm. In Section 5 we discuss the relationship between pAC and similar methods in the literature. Section 6 provides a summary of our empirical evaluation. In Section 7 we close with a discussion of these results. 2 BELIEF PROPAGATION IN BAYESIAN NETWORKS Kim and Pearl [8, 13] developed a polynomial time algorithm for singly connected Bayesian networks. The method is based on message passing, and there are two message types: causal messages, denoted by the symbol , are passed along the direction of the arcs in the DAG; diagnostic messages, denoted by the symbol , are ....

Jin H. Kim and Judea Pearl. A computational model for causal and diagnostic reasoning in inference systems. In Proceedings of the Eighth International Joint Conference on Artificial Intelligence, pages 190--193, 1983.

....below could have been used in our algorithms. The problem of computing posterior probabilities, exactly or approximately, in Bayesian networks is NP hard in general [5, 6] but in many cases the domain knowledge can be structured or simplified so that computation is feasible. 13 Kim and Pearl [24, 32] developed a polynomial algorithm for computing posterior probabilities for a special class of Bayesian network, known as polytrees. A polytree is a Bayesian network in which there is at most one undirected path between any two nodes. The algorithm works by message passing, each node passing ....

Jin H. Kim and Judea Pearl. A computational model for causal and diagnostic reasoning in inference systems. In Proceedings of the Eighth International Joint Conference on Artificial Intelligence, pages 190--193, 1983.

....inference and the certainty factor model are described in more detail in [1, 12] 3 Design of the belief network 3. 1 Basic notions of belief networks The formalism of belief networks offers an intuitively appealing approach for expressing inexact causal relationships between domain concepts [7, 20]. A belief network consists of two components [3] ffl A qualitative representation of the variables and relationships between the variables discerned in the domain, expressed by means of a directed acyclic graph G = V (G) A(G) where V (G) fV 1 ; V 2 ; V n g is a set of vertices, ....

....updated probability distribution after entering specific evidence into the belief network is called evidence propagation. Currently, there are two frequently applied algorithms for evidence propagation in use in belief networks. The first, and oldest, scheme originates from J.H. Kim and J. Pearl [7]. This scheme is only applicable in belief networks in which the causal graph has the form of a singly connected graph, i.e. a directed acyclic graph in which at most one path exists between any two vertices. The second scheme for evidence propagation originates from the work of S.L. Lauritzen and ....

J.H. Kim and J. Pearl, A computational model for causal and diagnostic reasoning in inference systems, in: Proceedings of the 8th International Joint Conference on Artificial Intelligence (Karlsruhe, West Germany, 1983) 190--193.

....from a medical textbook or by knowledge gathered from interviewing a doctor. 3.1 Heuristic, diagnostic reasoning There are several different languages known from the literature to formalize diagnostic reasoning. Examples of such languages, other than logic, are set theory and belief networks [33, 19]. In the present section, we focus on the logical representation of diagnostic reasoning in the spirit of MYCIN like rule based expert systems [4, 9] In this formalization, diagnostic reasoning is viewed as a deductive process instead of as an abductive process, the other frequently adopted view ....

J.H. Kim and J. Pearl, A computational model for causal and diagnostic reasoning in inference systems, Proceedings of the 8th International Joint Conference on Artificial Intelligence (Karlsruhe, West Germany, 1983) 190-193.

....which will allow to obtain good approximations to the solutions. It will be assumed that the graph is such that a simple probabilistic propagation is computationally feasible. 1 Introduction Propagation algorithms in graphical structures were first developed for the probabilistic case [19, 21, 30]. The essence of the efficiency obtained by using propagation comes from the factorization of the global probability distribution that can be deduced from the independence relationships expressed by the graphical structures. Soon, it was discovered that these independence relationships could be ....

Kim J., J. Pearl (1983) A computational model for causal and diagnostic reasoning in inference systems. Proceedings of the 8th IJCAI, Karlsruje, 190-203. 29

....computational efficiency for reasoning in Bayesian networks is the exploitation of specific independence relations to avoid the calculation of the full joint probability distribution. A variety of methods have been developed, each focusing on particular families of network topology. Kim and Pearl [2] have developed a distributed algorithm for solving singly connected networks (polytrees) In this scheme each node in the network obtains messages, representing all the evidence E, from each of its parents and child nodes. The singleconnectedness guarantees that the information in each message is ....

J. H.Kim, J. Pearl, "A Computational Model for Causal and Diagnostic Reasoning in Inference Engines", Proc. of the 8 th Int. Joint Conf. On Artif. Intell., August, 1983, Karlsruhe, Germany, pp.109-193.

....for Level of domestic consumption given this evidence. Also, you may be interested in the configuration of maximal probability given the evidence (also called the most probable explanation of the evidence) The growing interest in Bayesian networks is due to the inference algorithms developed [7, 9, 6, 11]. They exploit the structure such that the various calculations can be performed without actually calculating PR(U ) For the model in Figure 1, the largest table used has 448 configurations. Software for editing and running Bayesian networks is available commercially as well as free of charge ....

J. H. Kim and J. Pearl. A computational model for causal and diagnostic reasoning in inference systems. In Proceedings of the Eighth International Joint Conference on Artificial Intelligence, pages 190--193, 1983.

....emphasis on capturing the meaning of medical knowledge used in problem solving. These two frameworks are not necessarily incompatible as is demonstrated by the existence of probabilistic belief networks and in uence diagrams, where uncertainty and structure is combined in a unifying framework [Kim Pearl, 1983; Pearl, 1988; Shachter, 1986] Nevertheless, much of the structure of medical knowledge cannot be captured in terms of decision theory simply because probability theory is not suciently expressive for representing various kinds of semantic relationships where there are ways of handling ....

Kim, JH and Pearl, J, 1983. \A computational model for causal and diagnostic reasoning in inference systems", Proceedings of the 8th International Joint Conference on Articial Intelligence, pp 190-193, Karlsruhe.

....of Inference Mechanisms for Bayesian Belief Networks An efficient propagation of belief on Bayesian Networks has been originally proposed by J. Pearl (Pearl, 1982, Pearl, 1986a) I n his work, Pearl describes an efficient updating scheme for trees and, to a lesser extent, 39 for poly trees (Kim and Pearl, 1983, Pearl, 1986b, Pearl, 1988a) However, as the graph complexity increases from trees to poly trees to general graphs, so does the computational complexity. The complexity for trees is O(n 2 ) where n is the number of values per node in the tree. The complexity for poly trees is O(K m ) where ....

J. H. Kim and J. Pearl. A computational model for causal and diagnostic reasoning in inference engines. In Proc. 8th. Intern. Joint Conf. on Artificial Intelligence, pages 190--193, Karlsruhe, Germany, 1983.

....way of exploiting the nested junction trees technique to achieve such reductions. The usefulness of the method is emphasized through a thorough empirical evaluation involving ten large real world Bayesian networks and both the Hugin and the ShaferShenoy inference algorithms. 1. Introduction Kim and Pearl (1983) first formulated inference in a Bayesian network through message passing, but for singly connected networks only. Later this was extended to multiply connected networks, where the messages are passed in a junction tree (or join tree) corresponding to the network (Lauritzen and Spiegelhalter, ....

Kim, J. H. and Pearl, J. (1983) A computational model for causal and diagnostic reasoning in inference systems. In Proceedings of the Eighth International Joint Conference on Artificial Intelligence, pp. 190--193.

....that the tables are intractable for the computer. In this section we give a couple of propagation methods which may help overcome this problem. The first method trades space for time, and the second method relaxes the accuracy. 4. 1 Conditioning The first propagation method for Bayesian networks (Kim and Pearl 1983) only works for Bayesian networks where the DAG has no undirected loops (so called polytrees) and for general Bayesian networks Pearl proposed a method of transforming any Bayesian network into a set of polytrees (Pearl 1986) The method is called conditioning, and we shall briefly describe it ....

Kim, J. H. and Pearl, J.: 1983, A computational model for causal and diagnostic reasoning in inference systems, Proceedings of the Eighth International Joint Conference on Artificial Intelligence, American Association for Artificial Intelligence, pp. 190--193.

....specific causal independence models in simplifying knowledge acquisition (Henrion, 1987; Olesen et al. 1989; Olesen Andreassen, 1993) Heckerman (1993) was the first to formalize the general concept of causal independence. The formalization was later refined by Heckerman and Breese (1994) Kim and Pearl (1983) showed how the use of noisy OR gate can speed up inference in a special kind of BNs known as polytrees; D Ambrosio (1994, 1995) showed the same for two level BNs with binary variables. For general BNs, Olesen et al. 1989) and Heckerman (1993) proposed two ways of using causal independencies to ....

Kim, J., & Pearl, J. (1983). A computational model for causal and diagnostic reasoning in inference engines. In Proc. of the Eighth International Joint Conference on Artificial Intelligence, pp.

No context found.

J. H. Kim and J. Peal. A computational model for causal and diagnostic reasoning in inference systems. In Proceedings of the 8th International Joint Conference on Artificial Intelligence, pages 190--193. 1983.

No context found.

J.H. Kim and J. Pearl. A computational model for causal and diagnostic reasoning in inference engines. In Proceedings 8th IJCAI, pages 190--193. International Joint Conferences on Artificial Intelligence, Karlsruhe, West Germany, 1983.

No context found.

J. Kim and J. Pearl (1983), A computational model for causal and diagnostic reasoning in inference engines, in Proceedings of the Eigth International Joint Conference on Artificial Intelligence, Karlsruhe, Germany, pp. 190-193.

No context found.

Kim, J. and Pearl, J. (1983). A Computational Model for Causal and Diagnostic Reasoning in Inference Engines, Proceedings of the Eighth International Joint Conference on Artificial Inteligence, pp. 190-193.

No context found.

Jin H. Kim and Judea Pearl. A Computational Model for Causal and Diagnostic Reasoning in Inference Systems. In Alan Bundy, editor, Proceedings of the 8th IJCAI, Karlsruhe, Germany, August 1983. William Kaufmann.

No context found.

J. H. Kim and J. Pearl. A computational model for causal and diagnostic reasoning in inference engines. In Proceedings of the 8th International Joint Conference on Artificial Intelligence, Karlsruhe, West Germany, pages 190--193. International Joint Conference on Artificial Intelligence, August 1983.

No context found.

J. Kim and J. Pearl (1983), A computational model for causal and diagnostic reasoning in inference engines, in Proceedings of the Eighth International Joint Conference on Artificial Intelligence, Karlsruhe, Germany, pp. 190-193.

No context found.

J. Kim and J. Pearl (1983), A computational model for causal and diagnostic reasoning in inference engines, in Proceedings of the Eighth International Joint Conference on Artificial Intelligence, Karlsruhe, Germany, pp. 190-193.

No context found.

J. Kim and J. Pearl (1983), A computational model for causal and diagnostic reasoning in inference engines, in Proceedings of the Eighth International Joint Conference on Artificial Intelligence, Karlsruhe, Germany, pp. 190-193.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC