5436 articles – 2632 references  [version française]
Detailed view
12th International Conference on Artificial Intelligence in Education AIED 2005, Amsterdam : Pays-Bas (2005)
Attached file list to this document: 
PDF
Hakem-Khider-2005.pdf(130.6 KB)
DIANE, a diagnosis system for arithmetical problem solving
Khider Hakem1, Emmanuel Sander1, Jean-Marc Labat2, Jean-François Richard1
(2005)

We hereby describe DIANE an environment that aims at performing an automatic diagnosis on arithmetic problems depending on the productions of the learners. This work relies on results from cognitive psychology studies that insist on the fact that problem solving depends to a great extent on the construction of an adequate representation of the problem, which is highly constrained. DIANE allows large-scale experimentations and has the specificity of providing diagnosis at a very detailed level of precision, whether it concerns adequate or erroneous strategies, allowing one to analyze cognitive mechanisms involved in the solving process. The quality of the diagnosis module has been assessed and, concerning non verbal cues, 93.4% of the protocols were diagnosed in the same way as with manual analysis.
1:  CHART – Cognitions Humaine et ARTificielle
2:  L'UTES – L'Usage des Technologies Educatives en Sciences