KYBERNETISCHE PADAGOGIK AND INTELLIGENT TUTORING SYSTEMS

''Kybernetische Padagogik'' (cybernetic pedagogic) and research in ''I ntelligent tutoring systems (ITS)'' basically are based on the same id ea: to model the structure and process of instruction where the goal i s to develop a learner-adaptive, flexible and objective teaching progr ams. The theoretical background was developed by the cybernetic pedago gic. As H. Frank ([Frank70b], S. 260f) realized instructional processe s are kind of information processing. Therefore instruction can be des cribed as an feedback-driven information system. The behavior of this system is affected by six variables: the so called psychological struc ture (P) representing the learners initial state of knowledge as well as her learning abilities, the media (M) carrying the information, the teaching algorithm (B) representing the teaching strategies and abili ties the subject (L) and the teaching goals (Z) as well as the sociolo gical context (S) in which teaching fakes place (see figure 1, page 3) . In the context of cybernetic pedagogic the variables P, M, L, Z, S a re given B is searched for. Research in ITS founded a similar model of instruction. The basic elements of an ITS are the knowledge base (sim ilar to L) representing the knowledge the system has about the subject , the student model (similar to P) representing the knowledge which th e system diagnoses the learner has, the didactic model (similar to B) modulating teaching strategies and the user-interface (similar to M) ( see figure 2, page 4). An important improvement compared to cybernetic pedagogic is that the variables are modeled as dynamic variables whic h effect each other during the learning sessions. This became possible with the help of new methods in computing science, specially with the help of Al-methods. So if seems that the ideas of cybernetic pedagogi c could be put into practice due to new methods in artificial intellig ence. One weak point in ITS is the modeling of learning goals. But lea rning goals are important for the control of learning processes. I wil l give a short insight to my ideas to get over this weak point.