"Student modeling (SM) is recognized as one of the central problems in the area of Intelligent Tutoring Systems. Numerous SM approaches have been proposed and used with more or less success. Constraint-based modeling is a new approach, which has been used successfully in three tutors developed in our group. The approach is extremely efficient, and it overcomes many problems that other student modelling approaches suffer from. We present the advantages of CBM over other similar approaches, describe three constraint- based tutors and present our future research plans"
Mitrovic, Antonija, Michael Mayo, Pramuditha Suraweera and Brent Martin. “Constraint-Based Tutors: A Success Story.” IEA/AIE (2001).
1. SUMMARY
The paper describe three constraint-based tutors (SQL-tutors, CAPIT, and KERMIT) and the advantages of constraint-based modeling. Constraint-based modeling (CBM) represents domain knowledge as a set of state constraints each of which is in ordered pair (Cr, Cs). Violations of the pair indicate incorrect knowledge. Advantages of CBM includes: + Not requiring a runnable expert module/problem solver + Not requiring error libraries + Helping form the knowledge base of a runnable domain module + Computational simplicity + Not being sensitive to the radical strategy variability phenomenon.
2. STRENGTHS
The paper basically describes three use cases (systems) that involve CBM. Descriptions are very detailed and giving insights on the designs and operations of such systems.
3. WEAKNESSES
The paper is not easy to read due to the fact that its content deals with complex knowledge and the fact that the paper did not provide a good Background Knowledge section for readers who are not familiar with this knowledge domain. The writing of case study (3 products) does not appear to be straight to the point. The structure can be better organized to really highlight the advantages of CBM as we as the difficulties that CMB helps solving.
Image credit: Cisco.com
