Clark, P. (1988). Representing arguments as background knowledge for constraining generalisation. In EWSL'88 - Proc. 3rd European Working Session on Learning, pages 37--44.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... far as our quality based assessment model is concerned (cf. Section 6) the primitives of the underlying decision logic resemble those considered by Cohen s model of endorsements (Cohen, 1985) Such a quality based evaluation of competing hypotheses in a learning system has also been suggested by Clark (1988). However, the resolution strategy he proposes is not flexible enough as it crucially depends on a stronger than relation between sets of hypotheses that must be pre specified (some flexibility comes in through adaptive processes performed by a case based reasoner) In our scenario, however, we ....

Clark, P. (1988). Representing arguments as background knowledge for constraining generalisation. In EWSL'88 - Proc. 3rd European Working Session on Learning, pages 37--44.

....The immediate goal is to build up concept hypotheses for that concept and to evaluate and rank these tentative concept descriptions in quality based terms according to a credibility scale. Such a quality based evaluation of competing hypotheses in a learning system has also been suggested by Clark (1988). However, the resolution strategy he proposes is not flexible enough, as it crucially depends on a pre specified stronger than relation between sets of hypotheses. This requirement seems untenable for our application, since unrestricted combinations of quality assessments cannot be reasonably ....

Clark, P. (1988). Representing arguments as background knowledge for constraining generalisation. In EWSL'88 - Proc. of the 3rd European Working Session on Learning, pp. 37--44. London: Pitman.

.... by Davies and Russell in their theory of determinations [7] and used for helping to constrain inductive learning [46, 47] A specialised formalisation of this type of uncertain knowledge for representing arguments for and against a hypothesis has also been used for assisting in learning tasks [6]. 3.5 Exemplar Based Representations There has recently been an interest in exemplar based representations for machine learning (e.g. the systems Protos [1] Nexus [2] and by Kibler and Aha [20] A clear definition of exemplar based representation is hard to formulate, and the phrase refers ....

Clark, P. E. (1988). Representing arguments as background knowledge for constraining generalisation. In Proc. Third European Working Session on Learning (EWSL-88) London, D. Sleeman, Ed., Pitman, pp. 37--44.

....solutions over time results in improved performance. Additionally, this technique models the behaviour of much human problem solving and thus contributes towards its psychological validity and comprehensibility by an expert. Examples of recent work in case based reasoning are Chef [12] by Clark [8, 9], Protos [2] and Nexus [3] also see [1] A comparison of this approach with rule induction approaches is given in [7] While case based approaches are rising in popularity, it should be noted that many remain rather ad hoc with little theoretical underpinning. 3.5 Knowledge Compilation In ....

Clark, P. (1988). Representing arguments as background knowledge for constraining generalisation. In Proc. Third European Working Session on Learning (EWSL-88) London, D. Sleeman, Ed., Pitman, pp. 37--44.

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