Path: utzoo!attcan!uunet!zephyr.ens.tek.com!uw-beaver!milton!caesar.cs.montana.edu!uakari.primate.wisc.edu!samsung!cs.utexas.edu!mailrus!ncar!mephisto!prism!hms2!patty From: patty@hms2.gatech.edu (Patty Jones) Newsgroups: comp.cog-eng Subject: Re: Is this an ITS or what? Message-ID: <7100@hydra.gatech.EDU> Date: 15 Mar 90 18:35:25 GMT References: <1168fullerr@yvax.byu.edu> Sender: news@prism.gatech.EDU Lines: 81 fullerr@yvax.byu.edu writes: >I was wondering if some of you could give me some feedback on an idea. >I'm a psychology student who is trying to create an ad-hoc cognitive science >program from existing courses. One of these courses is an Intelligent >Tutoring class offered by our CS department. The model we are using in the >class is the one proposed in "Foundations of Intelligent Tutoring Systems" >by Polson and Richardson (1988). This model has an expert module, a student >module, a control module and a user interface. We have been argueing about >the difficulties of these components for two months now and we have all >decided that an ITS is impossible with the current technology. >I have come up with what looks like on the surface to be a simplification. >This simplification consists of pushing the control and student modules out >into the user. This would simplyify the system into an expert module with a >human-computer interface. The student would be able to learn a lesson module, >and then test himself by proposing the question and then answering his own >question thus bypassing the student module. The control module would be >overtaken by an interface that would allow the user to navigate the >information in any order. The real trick will be if I canget the metaphors >that I am working as a criterion knowledge representation working, then I will > be in better shape to help with the last two of those problems. >My question is then, is this an ITS or is it an expert system with a fancy >interface and a novice user? Any comments would be appreciated. >Rodney Fuller. It sounds to me like your proposed system allows the student to explore an expert knowledge base with the aid of some navigational tools in the interface (with a form of hypermedia?). This reminds me somewhat of the STEAMER system (Hollan, Hutchins, & Weitzman, 1984), which is an inspectable, dynamic, graphical simulation of a steam propulsion system. The basic idea is that by allowing the student to inspect the simulation while it's running, the student will build up the "correct mental model" of how the system works. Thus, the simulation serves as a form of "continuous explanation" to the student. I guess you could argue from this that your system *is* an ITS. BUT -- On the other hand, to me a major feature of an ITS is that it does model the student (in the student module) and uses this knowledge in conjunction with its pedagogical strategies (the "control module") to guide the interaction. Thus, an expert knowledge base is not enough (see Clancey, 1987, pp. 1-2 for a discussion as to why the MYCIN knowledge base needed to be supplemented with the GUIDON tutor. He says that you could just request explanations from MYCIN and learn from that, but that's inefficient and requires that the student ask exhaustive questions to learn everything in the knowledge base). Other comments: You said that the student would learn a "lesson module" -- who defines this and how is it defined? It looks like your proposed system will not know if the student answered her own questions correctly or not. If the student gets no feedback on performance, how can she learn? Plus, it also looks like the system won't know if the questions themselves are reasonable or valid. In summary, I think you have what could be called an "inspectable expert system". I think it will need more than navigational tools to support tutoring though. References: Clancey, W. J. (1987). _Knowledge-Based Tutoring: The GUIDON Program._ Cambridge, MA: MIT Press. Hollan, J. D., Hutchins, E. L. & Weitzman, L. M. (1984). STEAMER: An interactive, inspectable, simulation-based training system. _The AI Magazine_, Summer 1984. Reprinted in G. Kearsley, (Ed.), _Artificial Intelligence and Instruction: Applications and Methods._ Reading, MA: Addison-Wesley, 1987. patty Patty Jones Center for Human-Machine Systems Research School of Industrial and Systems Engineering Georgia Institute of Technology, Atlanta, GA 30332 UUCP: patty@chmsr.UUCP {backbones}!gatech!chmsr!patty INTERNET: patty@chmsr.gatech.edu