Instructional Authoring with Problem and Lesson Cases

This Small Business Innovative Research project uses analogical reasoning in machines to support analogical reasoning in humans. It was motivated by recent discoveries in psychology by Gentner and her colleagues indicating that, without explicit guidance,humans do a poor job of discerning useful analogies across cases, and that they often fail to retrieve useful analogous cases from memory when needed. The empirical evidence indicates that students do a better job of analogical encoding and futureknowledge transfer, if encouraged to construct analogies at learning time. In Phase-II, we propose to implement and evaluate the authoring tool and one example application Intelligent Tutoring System (ITS) as designed and prototyped during Phase-I. Ourdesign uses analogical reasoning technology to improve on conventional case-method instruction by assisting the author (military domain expert or instructor) in locating multiple analogous cases at lesson creation time. It then helps the author use thosecases in lessons that foster analogical encoding by students at runtime. We have obtained course material from National Defense University and Phase-II participation of military experts and educators to guide the design of the authoring tool and creationof an initial library of example military cases and lessons. The loose coupling of these two systems through a small set of MAC/FAC invocation calls by the RASCAL authoring and runtime systems, will lay the foundation for important future work in severalrelated areas: human dialog with agents that perform analogical reasoning, dialog agents that perform analogical reasoning as part of interpretation and generation of context-appropriate dialog contributions, advanced search engines that reason across acorpus of documents and across sequential queries indicating user needs, to name only a few.