Artificial intelligence helps advance personalized learning for South Carolina’s K-12 teachers and students

Artificial intelligence already recommends movies to viewers, filters spam out of inboxes and suggests friends on social media. Now the technology and some of its creators are going to work to help teachers and students in South Carolina’s K-12 education system.

Two recent projects have given researchers in Clemson University’s School of Computing  a chance to use their knowledge and experience in artificial intelligence to improve K-12 education in the state and eventually the nation. They are collaborating with researchers from the College of Education.

Nathan McNeese and Bart Knijnenburg and their students are helping create a “recommender system” similar to the one Netflix uses to suggest movies, except theirs will help teachers choose a path for professional development.

In a separate project, Knijnenburg and Kelly Caine and their students are helping develop AI-focused education modules for middle school students. The modules will teach math while also showing how AI’s invisible but powerful algorithms track them online.

McNeese and Knijnenburg are assistant professors of human-centered computing, and Caine is Dean’s Associate Professor of human-centered computing. 

Both projects underscore how artificial intelligence is continuing to evolve to meet a growing number of needs and becoming more deeply ingrained in everyday lives.

Clemson University’s researchers are helping take the technology to the next level with projects that vary from inspecting vehicles as they travel an assembly line to detecting malware in cameras and other devices too small for anti-virus software.

Intersection of disciplines

To Clemson officials, the research is further evidence that the most interesting work happens at the intersection of disciplines.

“The College of Education and the College of Engineering, Computing and Applied Sciences are coming together for interdisciplinary research that addresses some of the most critical needs in the state,” said George J. Petersen, founding dean of the College of Education. “Innovation is key to tackling education issues, and collaborations like these provide a perfect opportunity to put best-in-class innovation into action for the benefit of students, teachers, schools and communities.”

Principal investigators on both projects are faculty members in the College of Education. Jeff Marshall, associate dean of research and graduate programs in the college, is principal investigator on the recommender system research, and Nicole Bannister, associate professor of education, is principal investigator on the module research.

“This partnership brings together some of Clemson University’s best minds,” they said in a joint statement. “The knowledge and experience in artificial intelligence that Drs. McNeese, Knijnenburg and Caine and their students bring to these projects are crucial to making this research succeed. Our goal is for South Carolina’s students and teachers to ultimately benefit most. A better K-12 education system results in a healthier, happier and more prosperous society overall.”

‘Human-centered computing at its finest’

McNeese said that when the recommender system is ready for use, teachers will fill out a survey detailing their preferences and needs in professional development. Algorithms will process the data and provide the teachers with feedback.

Recommendations could include pursuing a master’s degree, a credential or an endorsement, all three of which are available through Clemson programs. 

“It’s human-centered computing at its finest,” McNeese said. “We’re going into the context, we’re talking to the user and we’re letting the user guide us with what they want and what they need. And then we’re taking that and doing the behind-the-scenes work to let them know their options. This has never been done for K-12 professional development to my knowledge.”

Knijnenburg said that he is helping create education modules for middle-school students who are online more than previous generations. Those students are leaving digital fingerprints about their interests, behavior and lives. 

“What I’m really hoping to convey is that these algorithms are making these inferences about their lives and their preferences,” he said. “You have to be careful in how you go about online, not just because other people can see it but because these algorithms are making predictions and inferences.”

Some insurers, for example, have explored using social media data to help determine premiums, Knijnenburg said.

Also as part of the project, researchers will use the core principles students learn in their math classes to explain how algorithms work and their consequences

 The modules will be offered online, allowing researchers to capture data, but will be used by students in the physical school classroom. Students will be able to access the modules through their school-issued electronic devices.

Addressing real-world challenges

Amy Apon, the C. Tycho Howle Director of the School of Computing, said the recommender system and the middle-school modules underscore how the school’s researchers are using AI to address real-world challenges.

“Researchers in the School of Computing are working across disciplines to apply artificial intelligence to a broad range of applications aimed at strengthening education, industry and individual well-being,” she said. “The research we do provides our students with a tremendous opportunity to work on the cutting edge of technology in a fast-growing field. Their experiences prepare them for the AI workforce of the future.”

Among those students are Christopher Flathmann and Reza Ghaiumy Anaraky, both Ph.D. students working on the recommender system.

Flathmann, the son of a special education teacher at Southside Christian School in Greenville, said he wants the recommender system to help get teachers the resources they need. He will have the opportunity to apply his expertise to the task.

“I am fairly good at looking at data and understanding patterns within data and seeing what we want and translating that into something that’s more understandable to people,” he said. “How do we make high-level, complicated computing things understandable to actual people? That’s what I do.”

Artificial intelligence generates a set of options, but it is equally important to communicate those options to individuals, Anaraky said.

Among the questions he is asking are: “How can we give individuals a sense of autonomy? How can we keep them in the loop and empower them to select an option that suits them best?”

Opportunities for undergraduates

The AI experience is also extending to undergraduates outside the School of Computing.

Kyra Derrick, a sophomore middle-level education major, said she recently transferred from Anderson University to Clemson University because she wanted the opportunity to study with Bannister.

Derrick, who is from Summerville, said she plans to interview middle school students about their feelings about math for the module research. 

“It’s going to be a great opportunity to just hear them,” she said. “Middle school students have a lot of thoughts, emotions and opinions. It’s going to be an amazing opportunity to let them feel heard and to be seen and for their opinion to be valued.”

Anand Gramopadhye, dean of the College of Engineering, Computing and Applied Sciences, said the two research projects highlight how working across disciplines yields maximum impact.

“These two projects are using artificial intelligence to advance personalized learning in the K-12 education system,” he said. “Both projects serve as examples of how Clemson University research at the intersection of disciplines  is helping solve the present and future challenges that society faces in our education system, while providing world-class educational experiences for students on campus and beyond.”