Learning processes in a class of knowledge‐based systems

Abstract

The paper is concerned with static plants described by a knowledge representation in the form of relations or logical formulas with unknown parameters. The learning algorithms consisting of step‐by‐step knowledge validation and updating have been presented for different versions. The theorems concerning the convergence of the learning processes have been given for the open‐loop and closed‐loop learning decision‐making system. The extension for the case with uncertain observations has been suggested. Simple examples illustrate the presented approach.

Citation

Bubnicki, Z. (2000), "Learning processes in a class of knowledge‐based systems", Kybernetes, Vol. 29 No. 7/8, pp. 1016-1028. https://doi.org/10.1108/03684920010342107