acm-header
Sign In

Communications of the ACM

ACM News

NC State Receives Grant To Study Artificial Intelligence In The Classroom


View as: Print Mobile App Share:
CyberPad-created graphical representation

Notebook computers with artificial intelligence-based software will allow fourth- and fifth-grade students to create graphical representations that model different scientific phenomena they've learned in the classroom.

Credit: North Carolina State University

While some students might dream of having a robot to help them with their homework, researchers at North Carolina State University are getting closer to making it happen.

NC State recently received a four-year, $3.5 million grant from the U.S. National Science Foundation (NSF) to explore using artificial intelligence as a learning tool inside fourth- and fifth-grade science classrooms. Artificial intelligence is the science of giving computers human-like abilities to understand, plan, communicate, perceive, etc.

The project includes creating "CyberPads" – computer notebooks with artificial intelligence-based software that allows a user to create graphical representations that model different scientific phenomena he or she learns in the classroom.

"Fourth- and fifth-grade science classes have a particular focus on physical and earth sciences. The CyberPads will support interactive scientific modeling for topics such as electricity, landforms, weather and climate," says James Lester, NC State professor of computer science and the project's principal investigator. "Students will actually be able to sketch out these different concepts using the program, and then see the models come to life with animation, sound and narration."

Assisting the students in using the CyberPads will be "PadMates" – or intelligent virtual tutors that support science learning through interactive scientific modeling. The PadMates will be able to recognize the understanding of the CyberPad user and interact with that student accordingly. For instance, a student struggling to grasp the concept of gravity would receive more detailed explanations and assistance than a student who quickly exhibits his or her understanding of the topic.

"Today's teachers are tasked with trying to provide focused instruction for each student in their classroom. And in a classroom with potentially 25 students at different points along a learning progression for any given science topic, you can see how extremely difficult – if not impossible – it is to give that sort of individualized attention," says Eric Wiebe, NC State associate professor of math, science and technology education, and co-principal investigator on the project. "For years, studies have shown the important role one-on-one tutoring plays in a child's understanding – so using 'virtual' tutors will provide teachers important assistance."

The program will initially be rolled out in 16 fourth- and fifth-grade science classrooms in North Carolina, to be followed by an additional 44 classrooms in Texas and California. As part of the program, professional development will be given to the teachers in order to prepare them to use CyberPads and PadMates in their classrooms as effective teaching tools. Researchers will study the students' problem-solving skills – such as strategy, thinking and collaboration – as well as the level of engagement in learning the science concepts before and after the use of CyberPads to see if, and how, the tools impact learning.

"Beyond supporting teaching overall science content knowledge, this system will also help students get creative in their problem-solving skills. This isn't just repeating back something they heard from their teacher – they'll actually be demonstrating an understanding of some very specific concepts," Lester says.

The project, "The Leonardo Project: An Intelligent Cyberlearning System for Interactive Scientific Modeling in Elementary Science Education," is funded by the NSF's Discovery Research K-12 program. In addition to Lester and Wiebe, the NC State project team includes Mike Carter, associate dean of the graduate school, and Bradford Mott, computer research scientist.


 

No entries found