Authoring Tools for Knowledge-Based Tutors

[ Overview | Tools | Current Titles | Journal Article | Behind the Scenes | Tom Murray's Home Page | Funding ]

Return to Tom Murray's home page

"Eon" is the name for our suite of authoring tools for building intelligent tutoring systems. Eon includes tools for authoring all aspects of intelligent tutors, including the learning environment, the domain knowledge, the teaching strategies, and the student model.  Funding and development for this project ended in 1997.   The Eon software, written in the prototype language SK8,  is unsupported prototype software that is not avaiable for distribution.

[picture of tool launcher]

Eon Overview:

Authoring tools such as HyperCard and Authorware are currently being used to create high quality multimedia educational titles. However, these titles tend to be fairly limited in their interactivity and ability to adapt to the student's needs. Intelligent Tutoring Systems (ITSs), which include such features as student profiling, coaching, and adaptive teaching strategies, are proving to be highly effective. However, authoring tools that support such advanced features are not yet available commercially, and ITSs must be build from scratch at considerable cost and effort. At the University of Massachusetts computer Science Department we have developed a suite of authoring tools called Eon which are specially tailored for creating these types of learning environments. The goal of Eon is to allow instructional designers to cost effectively build multimedia-rich cross-platform tutorials and learning environments with embedded intelligent instructional strategies.

The main thrust of the authoring tools work is in creating "knowledge based" tutorials, in which the instructional content is stored in a highly modular and re-usable form. Specifying how the content is to be sequence and presented to the student is done with generic, reusable "teaching strategies." Relationships between topics in the tutorial are specified using a concept networking tool. The system also maintains a "student model" of the mastery levels achieved for each topic, and the learning preferences for each student. Thus the content and style of instruction can be highly sensitive to the student's needs and learning history. For example, the author of a tutorial can create an adaptive teaching strategy specifying how and when hints are given to a student, and when examples should be presented. Presentation of a given topic can differ for different situations, for example, a high-achieving student may receive a summary with two examples, while a struggling student may get a more directive tutorial on that topic.

The highly modular format of the instructional material has a number of advantages:

  1. it facilitates creation of generic tutoring strategies and the inclusion of multiple strategies which are chosen based on the dynamically changing needs of the student;
  2. curriculum can be easily extended or modified to update information and theories, and new curriculum can be uploaded over the world wide web;
  3. teachers using the tutorials in their classroom can easily customize certain aspects of the tutorial to fit their needs (for example, by altering prerequisite relationships among topics or replacing pictures with ones more relevant to their classroom environment);
  4. we can easily build another tutorial while re-using the same teaching strategies;
  5. it facilitates a high degree of student control and initiative, since students can navigate to different points in the curriculum and since students can navigate to different points in the curriculum and ask for hints, explanations, etc., and these requests get processed using common formats.
Phase I of the Eon project is complete, and we are using our working prototype to build tutorials in chemistry, Japanese language learning, conceptual physics (statics), and in introductory thermodynamics.

Eon Tools

Tool Launcher
Provide access & information on other tools
  1. Document Browser

  2. Hierarchical access to all objects in the tutorial
  3. Topic Network Editor

  4. Define topics and the pedagogical relationships between them
  5. Tutoring Strategy Editor

  6. Specify, e.g. when to give hints or examples; how the topic net is traversed
  7. Interaction Editor

  8. Draw out the screens the student will see; content comes from templates or can be generated at run time
  9. Topic Contents Browser

  10. Link the topics (and their pedagogical levels) to the content
  11. Presentation Contents Browser

  12. Edit Contents for Interaction Templates
  13. Student Model Editor

  14. Relate student behavior to Topic & Lesson mastery
  15. Meta Strategy Editor

  16. Define when to change teaching strategies (e.g. from didactic to socratic)

Eon Titles - Intelligent Tutors Built with Eon

  1. Statics Tutor

  2. Introductory Physics
    Spiral teaching; misconception diagnosis; crane boom simulation
  3. Chemistry Workbench

  4. Simulated laboratory learning environment
    Solubility; chemical reactions and equations
  5. Bridging Analogies Tutor

  6. Socratic teaching strategy
    Derived from classroom-based cognitive studies
  7. Keigo Tutor

  8. "Honorifics" in Japanese language
    Rule-based expert system
  9. Thermodynamics Tutor

  10. Refrigeration; Air-conditioning

Eon Funding

The Eon project is supported in part by NSF and ARPA cooperative agreement # CDA-940860 and is part of The East/West Authoring Tools Group



ehaugsja@cs.umass.edu, tmurray@cs.umass.edu / 12/10/96