Poole, David, 1991. "Representing Diagnostic Knowledge for Probabilistic Horn Ab- duction", Proceedings, Twelfth International Joint Conference on Artificial Intelligence, Sydney, Australia, Vol. 2, pp. 1129-1135.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....to rule out the applicability of some hypotheses. Brewka [10] generalises this framework by blurring the distinction between facts and defaults and allowing any number of partitions of formulae which, for practical purposes, can be considered linearly ordered when constructing extensions. Poole [103] also extends this approach by introducing a probabilistic approach based on Bayes rule for determining the best explanation. As with other probabilistic approaches, he makes a number of assumptions to reduce the complexity of the calculations involved. 3.3.2 Abduction and Negation as Failure ....

David Poole. Representing diagnostic knowledge for probabilistic horn abduction. In Proceedings of the Twelfth International Joint Conference on Artificial Intelligence, pages 1129--1135, 1991.

....with regard to the notion of explanation. What accounts for an explanation What determines the quality of an explanation What is a best explanation What is a process of inference to the best explanation [48] Numerous works on abduction adopt the following general definition of explanation [18, 19, 46, 58, 59, 66, 76]. 1 Definition 1.1 Given a system S, and an observation Obs, an explanation of Obs in S is a formula expl(S; Obs) that satisfies the following conditions: i) S [ fexpl(S; Obs)g j= Obs; ii) expl(S; Obs) is consistent with S. Usually several refinements of this definition are considered. 1) ....

....of Obs such that an explanation ff is preferred to fi if EP (ff; Obs) EP (fi; Obs) and prob(ff) prob(fi) or EP (ff; Obs) EP (fi; Obs) and prob(ff) prob(fi) Unfortunately, this approach leaves many pairs of competing explanations incomparable. 3 An approach adopted by many researchers [8, 9, 31, 63, 64, 66, 79, 82], called Most Probable Explanation (MPE) in [63] or Maximum A Posteriory (MAP) assignment in [8, 9, 79] is selecting an explanation expl possessing a highest probability given all the available knowledge, that is, maximizing the posterior probability prob(expljS [Obs) This approach is based on ....

Poole, D., 1991, Representing diagnostic knowledge for probabilistic Horn abduction. In Proceedings, 12th International Joint Conference on Artificial Intelligence, Sydney, 1129-- 1135.

....a best abductive explanation to be defined probabilistically. Poole has proposed an account of probabilistic horn clause abduction. Further, he has shown a correspondence between his work and Pearl s belief nets. Although beyond the scope of this paper, the interested reader is referred to [55], 56] 3.1.1 Relating Closure to Abduction for Causal Frameworks The relationship between abduction, closure, consistency based reasoning and deduction was originally observed by Reiter [63] and described by Poole [51] and Console [9] 10] in the context of diagnostic reasoning with causal ....

D. Poole. Representing diagnostic knowledge for probabilistic horn abduction. In Proceedings of the Twelth International Joint Conference on Artificial Intelligence, pages 1129--1135, 1991.

....the strength of the causal link (which may then be used for preference purposes) whereas our preferences are independent of the strength of the causal link, and can be used to encode preferences on the basis of not only plausibility, but also danger, cost, urgency etc. The same remark holds for [12]. Although not in a diagnostic context, 13] has proposed a scheme for weighted abduction where literals in a theory (in our case: nodes in a causal net) are equipped with numerical weights, and minimisation of weights is used to select among competing explanations. However, assignment of the ....

....Extending our results beyond Horn Clause theories might be more problematic. Because we attach the preference predicates to causal links, we make very specific assumptions about the syntactic form of N . Although these assumptions are widely made in the literature on abductive diagnosis [2, 12], other formalisations of abduction deal with N as an arbitrary theory (e.g. 9] or the abductive case of [4] It is unclear how our approach could be used in theories that do not impose a particular syntactic form for individual causal links. A second extension involves an enlarged vocabulary ....

D. Poole, `Representing diagnostic knowledge for probabilistic Horn abduction', in IJCAI'91, pp. 1129--1135, (1991).

....urgently require explanation. This is the case in diagnostic applications, for example, where observations to be explained contradict our belief that a system is performing according to specification. The first two of these problems can be addressed using, for example, probabilistic information [29, 17, 46, 41]. We might simply require that an explanation render the observation sufficiently probable. Explanations might thus be nonmonotonic in the sense that ff may explain fi, but ff fl may not (e.g. P (fijff) may be sufficiently high while P (fijff fl) may not) For instance, it is highly likely ....

....explanations, there must exist some criteria for this choice. An obvious preference criterion on explanations is based on the likelihood of the explanations themselves. An agent should choose the most probable explanation relative to a given context. Such accounts are often found in diagnosis [46, 15] and most probable explanations are discussed by Pearl [41] In a more qualitative sense, one might require that adopted explanation(s) be among the most plausible. This view is advocated by Peirce (see Rescher [52] and Quine and Ullian [48] The notion of minimal diagnosis in the ....

David Poole. Representing diagnostic knowledge for probabilistic horn abduction. In Proceedings of the Twelfth International Joint Conference on Artificial Intelligence, pages 1129--1135, Sydney, 1991.

....while our approach allows the observation of any propositional formula) The same remarks apply to [10] Other related approaches include work in the domains of probabilistic abduction and model based diagnosis, temporal diagnosis, and also belief update. Probabilistic (and cost based) abduction ([12], 8] 5] attach probabilities to explanations from prior probabilities of hypotheses (faults in model based diag nosis) and generally an independence assumption amongst them. In the domain of temporal model based diagnosis, Friedrich and Lackinger [9] attach time intervals to fault ....

D. Poole. Representing diagnostic knowledge for probabilistic Horn abduction. In Proc. of the 12 th IJCAI, pages 1129--1135, 91.

....probabilities. The focus of this paper is on the implementation. 1 Introduction Probabilistic Horn Abduction [ Poole, 1991c; Poole, 1991b; Poole, 1992a ] is a framework for logic based abduction that incorporates probabilities with assumptions. It is being used as a framework for diagnosis [ Poole, 1991c ] that incorporates both pure Prolog and Bayesian Networks [ Pearl, 1988 ] as special cases [ Poole, 1991b ] This paper is about the relationship of probabilistic Horn abduction to logic programming. This simple extension to logic programming provides a wealth of new applications in diagnosis, ....

D. Poole. Representing diagnostic knowledge for probabilistic Horn abduction. In Proc. 12th International Joint Conf. on Artificial Intelligence, pages 1129--1135, Sydney, August 1991.

....This forms the basis of an anytime algorithm for estimating arbitrary conditional probabilities. The focus of this paper is on the implementation. Scholar, Canadian Institute for Advanced Research Logic Programming, Abduction and Probability 2 1 Introduction Probabilistic Horn Abduction [22, 21, 23] is a framework for logic based abduction that incorporates probabilities with assumptions. It is being used as a framework for diagnosis [22] that incorporates both pure Prolog and discrete Bayesian Networks [15] as special cases [21] This paper is about the relationship of probabilistic Horn ....

....Scholar, Canadian Institute for Advanced Research Logic Programming, Abduction and Probability 2 1 Introduction Probabilistic Horn Abduction [22, 21, 23] is a framework for logic based abduction that incorporates probabilities with assumptions. It is being used as a framework for diagnosis [22] that incorporates both pure Prolog and discrete Bayesian Networks [15] as special cases [21] This paper is about the relationship of probabilistic Horn abduction to logic programming. This simple extension to logic programming provides a wealth of new applications in diagnosis, recognition and ....

[Article contains additional citation context not shown here]

D. Poole. Representing diagnostic knowledge for probabilistic Horn abduction. In Proc. 12th International Joint Conf. on Artificial Intelligence, pages 1129--1135, Sydney, August 1991.

....and argues that it is better to invent new hypotheses to explain dependence rather than having to worry about dependence in the language. Scholar, Canadian Institute for Advanced Research. Probabilistic Horn abduction and Bayesian networks 2 1 Introduction Probabilistic Horn Abduction [48, 47] is a framework for logic based abduction that incorporates probabilities with assumptions. This is being used as a framework for diagnosis [48] that incorporates both pure Prolog [30] and Bayesian Networks [39] as special cases. This paper expands on [48, 47] and develops the formal underpinnings ....

....Canadian Institute for Advanced Research. Probabilistic Horn abduction and Bayesian networks 2 1 Introduction Probabilistic Horn Abduction [48, 47] is a framework for logic based abduction that incorporates probabilities with assumptions. This is being used as a framework for diagnosis [48] that incorporates both pure Prolog [30] and Bayesian Networks [39] as special cases. This paper expands on [48, 47] and develops the formal underpinnings of probabilistic Horn abduction, shows the strong relationships to other formalisms and argues that it is a good representation language in its ....

[Article contains additional citation context not shown here]

D. Poole. Representing diagnostic knowledge for probabilistic Horn abduction. In Proc. 12th International Joint Conf. on Artificial Intelligence, pages 1129--1135, Sydney, August 1991.

No context found.

Poole, David, 1991. "Representing Diagnostic Knowledge for Probabilistic Horn Ab- duction", Proceedings, Twelfth International Joint Conference on Artificial Intelligence, Sydney, Australia, Vol. 2, pp. 1129-1135.

No context found.

Poole, David, 1991. "Representing Diagnostic Knowledge for Probabilistic Horn Abduction ", Proceedings, Twelfth International Joint Conference on Artificial Intelligence, Sydney, Australia, Vol. 2, pp. 1129-1135.

No context found.

Poole, D., Representing Diagnostic Knowledge for Probabilistic Horn Abduction. Proc. IJCAI-91, Sydney, Australia, 1991.

No context found.

Poole, D., Representing Diagnostic Knowledge for Probabilistic Horn Abduction. Proc. IJCAI-91, Sydney, Australia, 1991.

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