How to CREATE Domain Content for Intelligent Tutoring Systems

Patten1, T., Metzger1, M., Hookway1, S., Sliva1, A., Lasser1, S., Wallace1, J., and Long2, R.

Generalized Intelligent Framework for Tutoring (GIFT) Users Symposium (GIFTSym3), Orlando, FL (June 2015)

A critical—but often overlooked—step in authoring intelligent tutoring systems is the initial retrieval and extraction of domain content. This step is increasingly challenging as the number and size of content resources grow rapidly. Enterprise repositories such as the Central Army Registry (CAR) and even the World Wide Web itself are valuable sources of domain content that should be easily exploited by authors of intelligent tutoring systems. Unfortunately, significant effort is currently required to transform this raw material into useful tutoring resources. In this paper we describe a framework for retrieving and extracting domain content that could serve as a useful addition to the Generalized Intelligent Framework for Tutoring (GIFT; Sottilare, Brawner, Goldberg, & Holden, 2012; Sottilare, Holden, Goldberg, & Brawner, 2013).

Content Retrieval and Extraction for Advanced Tutoring Environments (CREATE) is a prototype system that demonstrates several advanced technologies for turning raw materials into useful domain resources. CREATE uses the methods and standards of the Semantic Web to add rich semantic metadata that enables the material to be retrieved based on its meaning rather than on specific keywords. This semantic metadata also enables CREATE to recommend material based on the target learner’s occupational specialty. CREATE automatically creates rich pedagogical metadata including readability scores, Component Display Theory quadrants (Merrill & Wood, 1974), Bloom levels (Krathwohl, 2002), and Gagne events (Driscoll, 2000). An important innovation in CREATE is that this rich metadata is attached to document segments so that tutor authors can find and retrieve useful chunks of material rather than whole documents, which are typically too large for authoring purposes. Each of the key CREATE technologies will be described in subsequent sections below, followed by a discussion of the CREATE architecture and concluding remarks.

Charles River Analytics
US Army Research Laboratory


For More Information

To learn more or request a copy of a paper (if available), contact Terry Patten.

(Please include your name, address, organization, and the paper reference. Requests without this information will not be honored.)