Skip Navigation

University of Nebraska–Lincoln


Intelligent Learning Object Guide

The iLOG Team at Work

The iLOG Team at Work

What is iLOG?

Learning Objects (LOs) are self-contained lessons designed to instruct students on a variety of topics. The Sharable Content Object Reference Model (SCORM) provides a specification for LOs to help ensure interoperability with the wide variety of Learning Management System (LMS) players such as Blackboard and Moodle. SCORM-compliant LOs consist of a set of HTML pages containing the content of the LO and a manifest file containing metadata tags and other information necessary for the LMS player. The metadata tags are intended to help potential users find suitable learning objects and to assist in cataloguing existing LOs in repositories. For example a LO designed to explain basic algebra could have a metadata tag indicating its suitability for teenage students.
SCORM supports a wide variety of metadata tags, but they are all optional and the designer of the LO is responsible for supplying them. All too often, no metadata tags are provided, which makes it difficult to catalogue existing LOs. The iLOG project intends to employ artificial intelligence to automatically generate metadata tags for both existing and new LOs. This is accomplished through the use of an HTML wrapper which logs user interactions to a central database. Such user interactions include page navigation, HTML-generated events such as mouse clicks, and the use of the SCORM assessment component. Through the application of advanced agent-based artificial intelligence, iLOG automatically generates metadata tags for LOs based on user interactions.


The iLOG project was initiated in August of 2007 in collaboration with the center Nebraska Center for Research on Children, Youth, Families and Schools (CYFS). The first step was to design a wrapper capable of capturing user interactions with SCORM-compliant learning objects using JavaScript and PHP to communicate between the LOs and a MySQL database backend. The next ongoing step was the development of LOs based on introductory computer science concepts and their deployment to introductory-level computer science courses at the University of Nebraska, Lincoln. The alpha phase in fall 2007 consisted of 3 LOs deployed to a single course with approximately 30 students. After this phase, both the wrapper and the database backend were modified to improve the amount of detail in the captured user interactions. The beta phase in fall 2008 consisted of 8 LOs deployed to 4 courses with 271 students. This phase included a CYFS experiment on active versus passive exercise feedback that resulted in several published conference papers. The iLOG data was also used in a published conference paper on data mining. The full deployment phase in fall 2009 consisted of 16 LOs deployed to 5 courses with 360 students. This phase included another CYFS experiment on elaborative versus minimal exercise feedback resulting in several published conference papers. The iLOG data was also used in a journal paper on data mining (still under review) and in several unpublished papers on machine learning. After the full deployment, our focus has been refining LO content based on data analysis from the previous deployments resulting in several published conference papers. The second full deployment phase in fall 2010 will use these refined LOs.

The Functions and Searching LOs will be deployed in the spring semester of 2008.