College of Engineering News Room

New Grant Uses AI to Enhance Remote Instruction

By Brad Stager

Using online remote instruction to extend education beyond school campuses has been growing steadily since the turn of the 21st Century and recently has been increasingly deployed as a response to the COVID-19 global pandemic that has made classrooms risky places to be. 

Studying how students learn online and improving the effectiveness of remote instruction are topics that researchers from many disciplines are exploring. Among them is Autar Kaw, a mechanical engineering professor at the University of South Florida who is the principal investigator for a recently awarded $599,770 National Science Foundation grant to examine the effect of adaptive learning on flipped classrooms, where there is a mix of in-person and online instruction.

The research setting has changed since Kaw applied for the grant in December 2019 as USF has recently adopted a university-wide plan to start the 2020 fall semester with on-campus classes and shift to remote learning at the Thanksgiving break, creating a hybrid class model that differs from the flipped version of Numerical Methods Kaw would normally teach to undergraduates.

Autar Saw in the Classroom

Autar Kaw in the classroom.

The subject is one that Kaw routinely teaches and he has also published textbooks, supplemental websites and videos as learning aids for his students. An internet search for Numerical Methods Guy yields links to Kaw’s YouTube channel and blog. With a variety and depth of course materials to use, there is the opportunity to explore the flipped classroom model, which emphasizes learning material prior to classroom instruction, in new ways.

“It is more about what could be done at home should be done at home and what is better to do in class should be done in class,” says Kaw, who was recognized in 2012 as the U.S. Professor of the Year by the Council for Advancement and Support of Teaching and Carnegie Foundation for Teaching, an award recognizing undergraduate instructional excellence.

“It’s not about making a compromise, but taking the best of both worlds to make something better out of it.”

According to Kaw, the current research project is an opportunity to promote a personalized learning experience that can benefit students.

A major component of it combines student demographic data and class performance with predictive analytics to facilitate earlier intervention for students who may be struggling to learn course material. Kaw says this will lead to the development of updated course content and innovative tools to deliver it.

“We’ll be developing some new videos catered toward adaptive learning as well as simulations.”

Applying machine learning to education as a way to track a student’s progress and provide corrective options when needed is similar to continuous monitoring methods in other fields such as manufacturing and health, according to Ali Yalcin, an associate professor in the Department of Industrial and Management Systems, who is a co-principal investigator on the grant and working on the data end of the project.

“We’re incorporating machine learning to help students succeed” says Yalcin. “Many industries are moving forward with real time alerting and notifications, and the way they are doing this is through streaming analytics and machine learning.”

System and performance monitoring are most useful if meaningful responses to notifications can be provided in a timely fashion. Yalcin contrasts the project model with current routines that use results of testing that may not occur until well into a class term, losing valuable time to address knowledge gaps.

This NSF award follows an exploratory research project that examined ways to improve students’ pre-classroom instruction preparedness with adaptive lessons. It used an adaptive e-learning platform geared toward providing a personalized learning experience for each student and the results suggested to Kaw that the adaptive approach merited further consideration with the current grant.

The three-year grant is titled “Transforming Undergraduate Engineering Education through Adaptive Learning and Student Data Analytics” and the research team’s focus on providing timely adaptive interventions and  lessons, particularly among underrepresented minority groups, women and recipients of need-based funding like Pell Grants, helps the National Academy of Engineering meet its goal of "Advancing Personalized Learning," one of the organization’s 21st Century Grand Challenges.

Other co-principal investigators for the project are Andrew Scott of Alabama A&M University, Renee Clark of University of Pittsburgh, and Yingyan Lou of Arizona State University.

Kaw started with a plan for helping instructors of the course called Computational Methods about 30 years ago. This was on the heels of having developed software in 1990 with the help of an independent study student for a course in Composite Materials and that was distributed on a 3.5” diskette. The software, that has gone through several revisions, has been used by students from at least 100 universities worldwide.

As technology advanced, Kaw has developed new content and delivery methods. The impact that the COVID-19 global pandemic has had in disrupting existing educational models further motivates him to continue research in the field.

“I am approaching NSF for a supplemental grant where we can teach the whole course online because of COVID-19. There’s an opportunity here for education to go to the next level.”