By Greg Russell

University of Memphis faculty member Dr. Vasile Rus smiles after the mention of Transylvania’s most famous character, Dracula. Rus hails from the same region in Romania as did the real-life inspiration for Dracula, Vlad the Impaler, who wreaked havoc on invaders in the 1450s. But when it comes to discussing his own research, Rus quickly turns serious.

The associate professor of computer science, who hails from Cluj, Transylvania, has received a $1.65 million grant from the Institute for Education Sciences to develop an advanced intelligent tutoring system.

A research team that stretches over five U of M departments is being led by Rus and is seeking to develop the most effective intelligent tutoring system in existence.

Artificial intelligent tutoring systems make use of a computer program that provides customized instruction or feedback to students, often via software that includes a “talking head” on the computer that directly interacts with students during tutoring sessions.

Named DeepTutor, Rus describes his project as an advanced intelligent tutoring system that promotes deep learning of complex science topics through quality interaction and instruction between computer tutor and student.

“The system emphasizes quality of interaction during tutoring as a way of increasing tutoring systems' effectiveness,” Rus said. “This is a transformative concept in contrast with the conventional wisdom of the last decade or so that held the believe that as interactivity increases, the effectiveness of tutoring systems should keep increasing regardless.”

He said quality instruction in DeepTutor is assured by the use of “learning progressions,” a recently developed framework by the science education research community to describe students' natural paths to mastery.

Rus said DeepTutor is expected to provide better assessment techniques as well as an enhanced communication between the computer tutor and student that will result in a higher quality of interaction, and thus increased learning gains.

Dr. Vasile Rus

DeepTutor is a three-year project that began Sept 1 with a targeted completion date of fall 2013. It is expected to be one of the largest interdisciplinary projects at the U of M. Also involved are co-principal investigators Dr. Mark Conley, a professor in the College of Education; Dr. Anna Bargagliotti, an assistant professor of mathematics; Dr. Don Franceschetti, a professor of physics; and longtime artificial intelligence guru Dr. Art Graesser, professor of psychology.

So who will ultimately benefit from the DeepTutor artificial intelligence tutoring system?

“We initially will target three categories of students: college-bound 11^th- and 12^th-graders, vocational students and non-science majors who take conceptual physics at the college level,” Rus said.

But, he added, the project is built with scalability in mind, meaning that the eventual goal is to make it adaptable to students of all ages, from kindergarten to lifelong learners.

Rus said DeepTutor has two components that he believes will set it apart from other state-of-the-art dialogue-based tutoring systems.

“It relies on deep natural language and dialogue processing techniques and it also integrates the advanced educational framework of learning progressions, a recent development in science education research.

“DeepTutor will be the first tutoring system that integrates learning progressions as the educational backbone of tutoring. These features warrant high-quality tutoring which in turn will lead to highly effective instruction and learning.”

The website for the project is www.deeptutor.org.