Knowledge level model of a configurable learning system

Céline Rouveirol,Patrick Albert
DOI: https://doi.org/10.1007/3-540-58487-0_20
1994-01-01
Abstract:This paper presents the knowledge level model of a configurable learning system that follows a Generate and Test strategy. The knowledge level model makes explicit the elementary functionalities of the learning tool (referred to as learning operations), the control of the learning primitives (referred to as bias), and the different implementations of the learning primitives. The proposed model is based upon the inference structure formalism of KADS and will be used as an interface when interacting with the user. This explicit representation of learning operations and related bias will make the experimentation of different configurations of the proposed learning tool easier for a knowledge engineer developing a Knowledge Based application.
What problem does this paper attempt to address?