Artificial intelligence and tutoring systems: computational and cognitive approaches to the communication of knowledge
PROUST: Knowledge-based program understanding
ICSE '84 Proceedings of the 7th international conference on Software engineering
| Hi-index | 0.00 |
This article proposes a method of automatic diagnosis of the level of knowledge reached by the user. The technique that will be used for that consists in comparing the choices made by the user with those made by an expert system under the same conditions. This comparison will enable us to know by who - the system or the user - the most efficient choices have been made. The data thus gathered are of the utmost importance for both explanatory and tutoring systems. Through the GénéCom and Bateleur systems, a method aimed at guessing and evaluating the user's intentions will be presented. This method will enable the system to define the level of the user and as a consequence to know what has to be explained to him. It will also enable the system to find out its own shortcomings and to know what it has to learn.