Natalie Parde and her Teachable-Agent. Photo by William Snyder
First year graduate student Natalie Parde has become the third UNT College of Engineering student to earn a Graduate Research Fellowship from the National Science Foundation to support her research into natural language processing and human robot interaction. In the history of the Computer Science and Engineering department at UNT, Parde is the first student to earn this award. Parde’s grant is to fund her project titled “Building a Better Agent: The Next Generation of Teachable-Agent Technology,” in which she is programming a pedagogical robot, the teachable-agent, as a learning companion which takes the role of the student rather than the teacher.
“This will be a robot with educational and metacognitive goals,” Parde said. “I’m primarily looking at helping two different groups: students and the elderly. For students it can be used as a learning aid because you’re teaching things and refreshing the information you learned elsewhere. With the elderly, doing the same things help maintain their cognitive processes.”
Parde’s teachable agent is unique for its ability to communicate intelligently with its human instructor; early iterations of teachable-agent technology were more primitive and consisted of students creating visual knowledge maps using a computer, similar to laying out an outline for a paper. Though she is currently in the preliminary stages of her project, the ultimate goal will be a robot capable of carrying on a natural conversation with humans, much more powerful than the speech recognition technology used in iPhones to power Siri.
“I was doing a lot of reading about intelligent tutor systems, systems that are teaching things to people,” Parde said. “I thought, ‘what if we reversed that, and had people teach things to the system?’”
Parde initially chose to study journalism when she first arrived at college out of her interest in writing, language and communication. Finding herself unsatisfied with journalism she began to research other majors and found a better fit with computer science.
“I found that journalism just wasn’t right for me,” Parde said. “I wanted to do something that involved problem solving, so I signed up for Computer Science 1 and continued on from there. It was something on my own that I decided to investigate because I did like the idea of solving problems and using computers but I didn’t have any prior programming experience.”
By July of 2012 Parde was on a summer research team working with Dr. Gayatri Mehta on a project called UNTANGLED.
“[UNTANGLED] is basically a game that crowd sources mapping strategies," Parde said. "It's similar to a puzzle game: players untangle different blocks on a grid and try to get the highest score. What’s going on behind the scenes is that we’re taking the maps they created and using those to develop better ways to map hardware using mapping algorithms.”
Her first taste of research sparked an interest in Parde, and even today, rather abstract or explicit, all of Parde’s projects and research interests have applications to language and communication.
“That was when I started thinking that I would like to go on to grad school and get a Ph.D. I always really liked writing so even when I decided to study computer science I still wanted to do something that allowed me to work with language and communication,” Parde said.
Natalie Parde is currently working as a graduate research assistant in the Human Language Technologies (HiLT) laboratory at UNT’s Discovery Park. The lab is under the supervision of its director, Dr. Rodney Nielson, who holds a dual Ph.D. in Computer Science and Cognitive Science from the University of Colorado, Boulder. The work of the lab focuses its research on these two areas and perfecting ways for humans to better interact with the technology around them.
Category: Computer Science and Engineering