Abstract
A scheme is being developed for student modeling that can be helpful in pursuing links between neural network and symbolic approaches to training system development. The approach has two problems. First, it is not possible to have a certainty about whether a particular piece of knowledge is present in a student. A second problem is that microlevel detail may not be the only useful level for student modeling. The authors' approach to dealing with both of these problems is to layer a lattice of what they call student modeling variables on top of direct overlay information.

This publication has 0 references indexed in Scilit: