Automated Cognitive Diagnosis and Student Modeling

During the late 1970's and early 1980's, I became actively involved in the area of learning and problem solving. Following the lead of Anzai and Simon, I co-developed the Sage system, which learned search heuristics by decomposing the task into one of learning the conditions for selecting each operator, drawing training cases from decisions along solution paths.

In 1981, I moved to the Robotics Institute at CMU and was looking for funding. At that time, the Cognitive Science Program at ONR had funds for work on intelligent tutoring systems, and we hit on the idea of using Sage to learn not correct strategies but the errorful strategies of students. The common threads were Newell's problem-space hypothesis and the assumption that one could map each problem-space operator onto a separate rule in a production system.

The ensuing ONR project led to a new system, ACM, that adapted the methods in Sage to hypothesize paths taken by students and to induce the conditions on operators that generated those paths. The system constructed models of the more common bugs from VanLehn's analysis of subtraction errors, although some errorful strategies proved beyond the ability of the relational learning methods then available. Stellan Ohlsson and Stephanie Sage played a central role in the early stages, and Jim Wogulis contributed after I moved to Irvine. Stellan coined the term automated cognitive diagnosis, which better expressed our approach than the older term, student modeling.

Related Publications

Langley, P., Wogulis, J., & Ohlsson, S. (1990). Rules and principles in automated cognitive diagnosis. In N. Fredericksen, R. Glaser, A. Lesgold, & M. G. Shafto (Eds.), Diagnostic monitoring of skill and knowledge acquisition. Hillsdale, NJ: Lawrence Erlbaum.

Ohlsson, S., & Langley, P. (1986). Psychological evaluation of path hypotheses in cognitive diagnosis. In H. Mandl & A. Lesgold (Eds.), Learning issues for intelligent tutoring systems. New York: Springer.

Ohlsson, S., & Langley, P. (1985). Identifying solution paths in cognitive diagnosis (Technical Report No. CMU-RI-TR-85-2). Pittsburgh, PA: Carnegie-Mellon University, The Robotics Institute.

Langley, P., & Ohlsson, S. (1984). Automated cognitive modeling. Proceedings of the Fourth National Conference of the American Association for Artificial Intelligence (pp. 193-197). Austin, TX: Morgan Kaufmann.

Ohlsson, S., & Langley, P. (1984). Towards automatic discovery of simulation models. Proceedings of the European Conference on Artificial Intelligence. Pisa, Italy.

Langley, P., Ohlsson, S., & Sage, S. (1984). A machine learning approach to student modeling (Technical Report No. CMU-RI-TR-84-7). Pittsburgh, PA: Carnegie-Mellon University, The Robotics Institute.

Langley, P. (1983). Student modeling as strategy learning. Proceedings of the Fifth Conference of the Cognitive Science Society. Rochester, NY: Lawrence Erlbaum.

For more information, send electronic mail to

© 1997 Institute for the Study of Learning and Expertise. All rights reserved.