LINCOLN, Neb. (KMTV) – University of Nebraska-Lincoln researcher Mohammad Hasan is developing an app he hopes will help improve academic performance for STEM students.
With the use of machine learning and the application of AI Hasan “Messages from a future youwill provide students with targeted, real-time help that will improve their performance in STEM courses.
Using an avatar created from students’ photographs, users will be able to interact with their “future selves” and learn ways to improve their grades.
To learn more about the app, you can check out the full version below:
A University of Nebraska-Lincoln computer scientist is harnessing the power of artificial intelligence to help undergraduate STEM students improve their academic performance. His project will strengthen the pool of college graduates prepared for STEM jobs, the number of which in the United States is expected to increase by about 10% by 2030.
With a three-year, $600,000 grant from the National Science Foundation, Mohammad Hasan is developing a machine learning-based app, called Messages from a Future You, aimed at providing students with targeted, real-time interventions that improve their performance in STEM courses. Using the app, students can engage in a dialogue with their “future self”—an avatar derived from the student’s photograph—about how to improve their grade.
The app would be the first to use an artificial agent to provide personalized interventions that take into account the myriad of factors that affect a student’s final grade.
“Existing approaches primarily target academic improvement based solely on academic performance,” said Hasan, assistant professor of big data and artificial intelligence in the Department of Electrical and Computer Engineering. “But end-of-semester performance is not only influenced by academic activities during the semester. It is shaped by other factors: family background, socio-economic status, interactions with peers, interactions with the instructor, scientific identity, etc. »
The app would be a portable and cost-effective tool to combat a national STEM major attrition rate that hovers around 50%, primarily due to students’ poor academic performance. With seven years of experience teaching large introductory courses in Nebraska, Hasan had first-hand knowledge of why undergraduates can struggle at the start of a STEM major.
“Helping students in large classes isn’t easy, because you can’t really talk to everyone throughout the semester, and students wouldn’t come to you unless they were really in trouble” , did he declare. “And usually when they come, it’s at the very end of the semester, when it’s not really possible to help them in any meaningful way.”
Hasan started thinking about ways to improve student performance. Although change at the institutional level may hold the most power to move the needle, Hasan believed there were smaller scale and less expensive ways to help students.
He recognized the power of machine learning to power an app that would help students succeed: the app could “learn” about the student’s behavior and background, then use that data to predict future performance and provide advice on how to modify this trajectory.
Through the app, students reluctant to ask an instructor for help — those who are introverted or feel intimidated, for example — would have a way to access personalized assistance.
“It’s like a proxy for me,” he said. The proxy would take shape onscreen as a lookalike of a student from the future, which Hasan said would resonate more strongly than just text messages.
To create a pilot app, he collected data on the academic performance of about 300 Husker undergraduate students in a computer science course and created a grade prediction app. He pitched the pilot to a sophomore class.
The app has increased the number of successful students. But the results showed that despite being on a similar academic trajectory at a certain time and receiving the same messages from the app at those times, the students’ final grades weren’t the same.
The NSF project aims to identify the factors behind these variable results. Hasan’s hunch is that the one-dimensional nature of the pilot tool — its basis on academic data alone — painted an incomplete picture of a student’s story, missing key factors that would explain the different results.
To fill this gap, he teamed up with project co-investigator and former Husker researcher Bilal Khan, developer of the Open Dynamic Interaction network. Using the ODIN software platform, they collect nine dimensions of academic, social, and psychological data to train the model.
Hasan will then use a machine learning technique called clustering – the automatic clustering of data – to organize students’ trajectories into distinct “story types”.
Identifying these types of stories is the keystone of the project: by “knowing” which group a given student belongs to, the app will provide the right messages and advice.
“If a student falls into story type A, we will understand that the student falls into a particular behavioral pattern,” Hasan said. “By knowing this pattern, we can provide a more targeted intervention to the student.”
Co-researcher Neeta Kantamneni, associate professor and director of the university’s counseling psychology program, is leading efforts to develop the interventions. His team will develop messages to deploy to students at certain times, depending on how their story unfolds.
Interventions will reflect the type of advice provided during face-to-face consultations. Students can be encouraged to participate in mindfulness activities to reduce stress or worry, collaborate with peers to build a sense of community, or seek extra help during office hours.
A major benefit is that interventions will come to students in real time, providing them with help when they need it most. This is an advantage over in-person counseling, which often occurs after students are no longer satisfied with their major.
The app would also strengthen students’ access to help during a time of long waiting lists for in-person counseling, both at the university and nationally. Kantamneni is optimistic that virtual interventions could help close this gap.
“We don’t want people leaving an entire career field just because they had a tough class,” Kantamneni said. “That’s what we see sometimes. And my goal is that that doesn’t happen, especially for students who are underrepresented in those classes — so women, students of color, first-generation students.
The app’s design has potential beyond the STEM education field, Hasan said. Allowing people to “see” how behavior today influences tomorrow could be particularly salient in public health. Seeing yourself sick could be a powerful boost to following medical recommendations like wearing a mask, exercising, or getting more sleep.
Hasan also hopes the app will highlight the positive potential of smart machines.
“I know these days people are very worried about the role of AI,” he said. “But by using AI for social good, we can show that there’s this other side of AI, which isn’t for making money or just doing business. We can use the AI to improve the lives of students Within the university, it is a very useful complement to improve the quality of teaching.
Download our apps today for all our latest coverage.
Get the latest news and weather straight to your inbox.