Aspects of a Theory for Automated Student Modelling.
Abstract
This report documents some of our recent investigations into a theory for automatically inducing and using (structural) models of a student which explicate his reasoning strategies, his representation of procedural skills and his underlying misconceptions as manifested in his errors. The first chapter discusses a diagnostic model based on the concept of a procedural network - a network having many of the properties of the older style semantic networks but which captures both the intensional and extensional (or executable) aspects of procedural skills. These diagnostic models provide not only a technique for modelling the underlying or deep structure aspects of a procedural skill but they also suggest that an important criterion for modelling cognitive processes and their related knowledge representation is that of finding a natural way to account for all possible manifested errors in human performance of that skill. The second chapter describes a considerably more complex theory/technique for analyzing the problem solving trace or protocol of a student and then automatically synthesizing a model of his problem solving strategies as well as the motivations or plans that he used to guide him in his solution. This theory captures the subtle reasoning powers of a master tutor and as such acts as a powerful modelling technique of a learner which is needed for guiding our computer-based laboratory tutor as well as providing a new methodology for measuring how a student's problem solving performance is evolving. This theory also forms a cornerstone for building information processing models of master tutors.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1977
- Accession Number
- ADA043271
Entities
People
- Bill Huggins
- Catherine Hausmann
- Ira Goldstein
- John Seely Brown
- Richard R. Burton
Organizations
- BBN Technologies