I. Weber and B. Tausend. A three-tiered confidence model for revising logical theories. In Stefan Wrobel, editor, Proc. Fourth Int. Workshop on Inductive Logic Programming (ILP-94), pages 391 -- 402, Schloß Birlinghoven, 53754 Sankt Augustin, Germany, 1994. GMD (German Natl. Research Center for Computer Science). Order from teuber @gmd.de.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....class, continuous degrees of confidence in each statement (e.g. identified with the ration of positive to negative examples covered by a clause) The system then selects to modify those clauses that are in the lowest confidence class and together have the minimal degree of confidence. In Miles [50], the confidence model was extended by a third tier which allows the introduction of other heuristic measures of confidence such as indispensability (number of positive examples that will become unprovable if a clause is modified) or the introduction of user interaction. Krt uses declarative ....

....In its simple form, the revision point or result is selected simply by asking the user. This is offered as an option in the Krt system [55] where the user can select among the possible minimal base revisions computed and heuristically ordered by the system, and in the revision module of Miles [50]. The Mobal system also includes procedures for fixing violations of integrity constraints (generalized theory revision task) These procedures also simply ask the user which premise instances to delete or which conclusion instances to add (where deletion may in turn give rise to theory ....

[Article contains additional citation context not shown here]

I. Weber and B. Tausend. A three-tiered confidence model for revising logical theories. In Stefan Wrobel, editor, Proc. Fourth Int. Workshop on Inductive Logic Programming (ILP-94), pages 391 -- 402, Schloß Birlinghoven, 53754 Sankt Augustin, Germany, 1994. GMD (German Natl. Research Center for Computer Science). Order from teuber @gmd.de.

....such a theory exists) Hence the problem is interesting only because of the inductive bias consisting of restricting revision to a subset of the input theory s components. Such a restriction is often used as an initial restriction on the revision process (Koppel et al. 1994; Saitta et al. 1993; Weber Tausend, 1994), but also can arise during the process of revision when an algorithm performs an unretractable revision. Thus the tools developed in this paper for analyzing the theory patching problem may be applied to many theory revision algorithms extant in the literature. We consider both propositional and ....

Weber, I., & Tausend, B. (1994). A three-tiered confidence model for revising logical theories.

....deeper within the proofs. Therefore, the more sophisticated minimal base revision approach [Wro93b] selects among the minimal sets of clauses required to cut off all successful proofs of negative examples. In MOBAL [MWKE93] the selection is controlled by a two tiered confidence model. In MILES [WT94], this model has been extended to a flexible three tiered confidence model. For investigating argument choice mechanisms in the bias shift setting, it suffices to consider the case that only one over general clause is to be specialized with a new predicate. In case it is necessary to specialize ....

I. Weber and B. Tausend. A three--tiered confidence model for revising logical theories. In Proc. of ILP-94, 1994.

No context found.

Weber I., Tausend B.: A Three-Tiered Confidence Model for Revising Logical Theories, in Proceedings of the Fourth International Workshop on Inductive Logic Programming (ILP-94), GMD-Studien Nr. 237, pp.391-402, 1994.

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