Feldman R., Segre A. and Koppel M., "Incremental Refinement of Approximate Domain Theories". Proceedings of the 8th International Conference on Machine Learning, 500-504, Evanston,IL,1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....17, 18, 19, 23] is not a brand new search algorithm, but, rather, a more informative search structure, called LT Tree (Learned Theory Tree) that enables the system to learn a structured knowledge base limiting the complexity of the search process. The LT Tree is not a new concept: Feldman et al. [7] addressed the problem of refining a domain theory, by proposing an incremental algorithm that modifies a special data structure, called DT tree: a DT p tree is an AND OR tree representing a full expansion of the domain theory rooted at predicate p. WHY [22] uses an AND OR graph representation of ....

Feldman, R., Segre, A., and Koppel, M. "Incremental Refinement of Approximate Domain Theories", Proc. of the Eigth International Workshop on Machine Learning, Evanston, IL, 500-504, 1991.

....theory. They then hill climb up the path from the selected revision point to the root of the theory, evaluating appropriate revisions to each node in the path, and selecting the one that is the simplest. This adds to the computational cost of these techniques. Probabilistic theory revision (PTR) (Feldman, Segre, Koppel, 1991; Koppel et al. 1994) takes a different approach to revising propositional Horn clause theories. This technique uses probabilities to express confidence in the correctness of different portions of the theory, which are then used to locate faults. PTR converts the logical theory into an AND OR ....

Feldman, R., Segre, A., & Koppel, M. (1991). Incremental refinement of approximate domain theories. In Proceedings of the Eighth International Workshop on Machine Learning, pp.

....made to the theory during previous batches (perhaps based on very little evidence) are no more suspect than rules from the initial theory. An incremental theory revision system needs ways of representing its confidence in various portions of the theory and then weighting changes accordingly [ Feldman et al. 1991 ] Either s results need to be compared to other theory revision systems [ Towell and Shavlik, 1991; Ginsberg, 1990 ] used in incremental batch mode. The ability of incremental batch theory refinement to track concept drift also needs to be explored. Of course, since a full memory system ....

R. Feldman, A. Segre, and M. Koppel. Incremental refinement of approximate domain theories. In Proceedings of the Eighth International Workshop on Machine Learning, pages 500--504, Evanston, IL, June 1991.

....bias language. The purpose of such a language is to allow the user to declare his own bias in a simple and structured way, that is to express the conditions placed on the domain knowledge for a given revision operator to be applied. This language extends and generalizes the work reported in [7,8] by considering a larger family of conditions. In the absence of a bias scheme the system will use a predefined cost scheme (each revision operator has a cost associated with its application) and suggest revisions that have a minimal cost. 3.1 Typical Situations In Which An Explicit Bias Is ....

Feldman R., Segre A. and Koppel M., "Incremental Refinement of Approximate Domain Theories". Proceedings of the 8th International Conference on Machine Learning, 500-504, Evanston,IL,1991.

No context found.

Feldman R., Segre A. and Koppel M., "Incremental Refinement of Approximate Domain Theories". Proceedings of the 8th International Workshop on Machine Learning, 500-504, Evanston,IL,1991.

No context found.

Feldman, R., Segre, A., and Koppel, M. "Incremental Refinement of Approximate Domain Theories", Proc. of the Eigth International Workshop on Machine Learning, Evanston, IL, 500-504, 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