Y. Kono, M. Ikeda, and R. Mizognchi, "THEMIS: a nonmonotonic inductive student modeling system," Journal of trtificial Intelligence in Education, vol. 5(3), pp. 371-413, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....procedure is automatically updated each time students interact with the systerm According to Sison and Shimura [5] machine learning or machine learning like techniques have so far been used in two areas of student modeling research. First, many systems, such as DEBUGGY [6] ACM [7] THEMIS [8] and ASSERT [9] have used methods inherited from the area of machine learning in order to induct a single, consistent student model from multiple observed student behaviors. Furthermore, machine learning or machine learning like techniques have also been utilized for the purpose of automatically ....

Y. Kono, M. Ikeda, and R. Mizognchi, "THEMIS: a nonmonotonic inductive student modeling system," Journal of trtificial Intelligence in Education, vol. 5(3), pp. 371-413, 1994.

....initialization procedure is automatically updated each time a student interacts with the system. According to Sison and Shimura [7] machine leaming or machine leaming like techniques have so far been used in two areas of student modeling research. First, many systems, such as DEBUGGY [8] THEMIS [9] and ASSERT [10] have used methods inherited from the area of machine leaming in order to induct a single, consistent student model from multiple observed student behaviors. Furthermore, machine learning or machine leaming like techniques have also been utilized for the purpose of automatically ....

Y. Kono, M. Ikeda, and R. Mizoguchi, "THEMIS: a nonmonotonic inductive student modeling system," Journal of Artificial Intelligence in Education, vol. 5, pp. 371-413, 1994.

....Instructional methods : Didactic, Inquiry and Discovery. We conclude by presenting conclusions and further work strands. Introduction Current Intelligent Tutoring System (ITS) design and implementation is based mainly on the Instructivist Educational Paradigm (Reeves, 1994; Keegan, 1993; Kono, Ikeda Mizoguchi, 1994; Cooper, 1993) and an associated tool, Instructional Systems Design (ISD) Capell, 1993 ; Zhou, 1996; Dick, 1991) For the most part, ISD is related to Behaviorism, as teaching learning results are directly evaluated through inspection of tangible data and conduct. Therefore, ISD calls for ....

....of the rationale behind the completion of these modules lies within question and answer evaluations and or reinforcements. Pierre Dillenbourg states in (Dillenbourg, 1996) that a major contribution of the AI in Education community has been the shift of focus from the search for answers method (Kono, Ikeda Mizoguchi, 1994), to a focus on the reasoning process of the learner, that is, to a cognitive approach. Although we agree with him, we believe much work remains to be done as to the integration of cognitive theory and the design of computer based instruction. ISD and IG s lack the semantic power to cover a ....

Kono, Yasuyuki; Ikeda, Mitsuru; Mizoguchi, Riichiro (1994). THEMIS: A Nonmonotonic Inductive Student Modeling System. JI.

.... = Obs (Michalski, 1983; Michalski, 1991) Machine Learning to maintain Student Models The first approach used for integrating MLT within ITS is related to maintaining the Student Model (Gilmore Self, 1988; Woolf Murray, 1994) In PIXIE (Sleeman, 1983) ACM (Langley Ohlson, 1984) THEMIS (Kono, 1993) or ELECTRE (Palis, Caillot, Cauzimille Marmche, Laurire, Mathieu, 1986) the Student Model is viewed as procedures for problem solving represented in a production rules formalism. In such systems, the Obs are observations of learner s correct and incorrect results, the background knowledge BK ....

Kono, Y. (1993). THEMIS: A nonmonotonic inductive student modeling system (AI Technical Report No. AI-TR-93-3). Osaka University.

....[21] Student modeling is the process of creating and maintaining students models. Acquiring these models from observable behavior is a hard task because is based on guesses about the learner [12] A lot of work has be done with artificial intelligence techniques to model student s reasoning [6] [4] 1] and with Bayesian networks to model student s behavior in a probabilistic way [3] Fuzzy logic techniques have been used to improve the performance of an ITS due to their ability to handle imprecise information, such as student s actions, and to provide human descriptions of knowledge and ....

Kono, Y. Ikeda, M. and Mizoguchi, R. (1994). THEMIS: A Nonmonotonic Inductive Student Modeling System. Journal of Artificial Intelligence in Education, 5(3), p. 371413

....are insufficient to characterize errors. Behavior recognition is more important when dealing with complex behavior (e.g. programs) than when the behavior is, say, a single number. Hence, DEBUGGY (Burton, 1982) and ACM (Langley and Ohlsson, 1984) which deal with numbers in subtraction, and THEMIS (Kono et al. 1994) and ASSERT (Baffes Mooney, 1996) which deal with labeled propositions, perform relatively trivial behavior recognition. The bulk of their work is in the induction of the student model from a set of noncomplex behavior. 3 Definition 3 Behavior characterization is the process of differentiating ....

....one end and transformational induction on the other. 1. Synthetic model induction involves selecting and combining model components and subcomponents, and then iteratively making modifications on this set until it finally explains student behavior. This is the approach of DEBUGGY, ACM, and THEMIS (Kono et al. 1994), which we shall take up later. 2. Transformational model induction, on the other hand, involves modifying an existing, usually ideal, model so that the final model explains student behavior. This is the approach taken by systems which use transformation operators for recognition and ....

[Article contains additional citation context not shown here]

Kono, Y., Ikeda, M., & Mizoguchi, R. (1994). THEMIS: A nonmonotonic inductive student modeling system. Jl. of Artificial Intelligence in Education, 5(3):371-413.

....the student model is built without considering the different moments at which data about the student have been collected. This hypothesis makes it possible to simplify the modeling process greatly, but unfortunately it seems too far removed from a realistic view (cf. Huang et al. 1991; Kono et al. 1994; Paiva et al. 1994; Paiva and Self, 1994; Errico, 1996) Let us consider the dialogue excerpt given in Figure 1, taken from the classic paper of Stevens and Collins (1982, p. 18) in the domain of meteorology, where t 7 , t 8 , etc. refer to different instants in the dialogue. This dialogue ....

Kono Y., Ikeda, M. and Mizoguchi, R. (1994). THEMIS: A nonmonotonic inductive student modeling system. Journal of AI and Education 5:371-413.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC