Rice University's OpenStax Tutor Tackles Personalized Learning

Machine learning is key to scaling personalized learning, presenters from Rice and Duke argue at SXSWEdu.
Although these cognitive principles could be applied without the use of technology, the benefit of an online learning system is the volume of data it can collect and drive back into future improvements, Butler said. Early results from OpenStax Tutor show that the boost in knowledge retention, just by driving students to retrieve and use their knowledge more frequently, is worth about a half letter grade in performance improvement, he said.

Attempts to personalize the learning experience are not new. Their downfall in most cases has been that they are too expensive and labor intensive, requiring the hand coding of rules assembled by lectures in each discipline, Butler said. OpenStax Tutor is trying to break that cycle by applying machine learning techniques to the problem of analyzing student performance and making recommendations for further study, he said. "We need to make this technology cheaper and easier to apply."

Although the idea that a computer could figure out study recommendations for students without human intervention might be foreign to educators, it's exactly the sort of thing Netflix and Amazon are doing when they recommend additional products you might enjoy based on your past search and purchase activity, Baraniuk said. "Tremendous progress has been made on so-called recommender systems."

OpenStax Tutor is designed so that individual students decide whom their data will be shared with, making it possible for them to be working with multiple instructors and learning coaches on the same material. Still, he recognizes the continuing need to focus on issues of student data privacy that will be "as vexing as anything having to do with electronic medical records."

Currently in beta, OpenStax Tutor is being made available as a free Web resource, although in the future there could be a charge to institutions that implement it on a large scale, Baraniuk said. So far, it works best with multiple-choice assessments, although he hopes that technologies for assessing free-form text will advance rapidly in the coming years. The application is designed for use with other open educational resources, such as those from Connexions or Quadbase assessment quizzes. However, individual professors can assemble any assortment of Web-accessible resources, including proprietary ones, to create homework assignments or digital study guides for their students, Baraniuk said.

"My feeling is the more open your tools, the more users your system is going to have and the more data you get to work with," Baraniuk said.

Follow David F. Carr on Twitter @davidfcarr or Google+.

Attend Interop Las Vegas, May 6-10, and attend the most thorough training on Apple Deployment at the NEW Mac & iOS IT Conference. Use Priority Code DIPR03 by March 9 to save up to $500 off the price of Conference Passes. Join us in Las Vegas for access to 125+ workshops and conference classes, 350+ exhibiting companies, and the latest technology. Register for Interop today!

Editor's Choice
Sara Peters, Editor-in-Chief, InformationWeek / Network Computing
John Edwards, Technology Journalist & Author
John Edwards, Technology Journalist & Author
James M. Connolly, Contributing Editor and Writer