Research

USC Rossier Assistant Professor Stephen Aguilar leads team to develop AI-enhanced dashboard for instructors and students

USC Rossier, USC Institute for Creative Technologies and USC Information Sciences Institute collaboration secured $4.5 million U.S. Army Research Office grant.

By Ellen Evaristo Published on

This fall, USC Rossier Assistant Professor Stephen Aguilar began work on the AI-Enhanced Dashboards (AID) project, one of the five inaugural tracks for the Artificial Intelligence Research Center of Excellence for Education (AIRCOEE). AIRCOEE is a two-year $4.5 million dollar research contract through the U.S. Army Research Office housed within the USC Institute for Creative Technologies (ICT) and led by Benjamin Nye (PI) and Bill Swartout, Chief Technology Officer of ICT. “The AIRCOEE effort is about AI from two perspectives: first, we are using AI to build enhanced tools and systems for education; second, the content area we are focusing on is itself AI, to help prepare the workforce for the jobs of the 21st century,” said Bill Swartout.

Professor Aguilar leads a team to create an AI-enhanced dashboard for instructors and students. The resulting set of dashboards will provide AIRCOEE with actionable insights to help instructors engage and support students in the classroom and during periods when they need to self-pace their learning.

With a background in learning analytics, Aguilar will take what is known about dashboard design and instructor insights to determine when to provide assistance to students who may need additional support. “We will use data from their learning management system and create a series of dashboards that are useful for instructors,” said Aguilar. “We aim to support the instructor’s pedagogy and provide information they'll need in order to be effective in their roles.”

Part of the research is the psychology of relevant user information. The team asked: how do you design a dashboard to be user friendly so individuals can easily locate information? Another part is determining what information to display. Salient information for instructors leads to better student outcomes.

“The idea is to provide sufficient actionable information to the user, and in this case the instructor. They could quickly learn about their students as a whole, learn about a particular student and then determine subsequent courses of action to help the student do better,” said Aguilar.

USC ICT has a history of developing cutting-edge technology for the military, such as conversational AI agents, virtual reality headsets and high-fidelity facial scanning. Many of these innovations have been transitioned to civilian use, through collaborations with industry, academic partners and even Hollywood. In addition to leading the overall AIRCOEE effort, ICT will develop the AI-Enhanced Dashboard software based on the designs led by USC Rossier.

On the technology, Nye stated, “We’re looking to make a dashboard that goes beyond just describing ‘what happened’ and instead suggests `what could happen next.’ Predictive models are part of this, such as early warnings about struggling students. However, an intelligent dashboard also needs to consider an instructor as an active partner: what interventions would an instructor consider for this student, and what information do they need to take action? This research also complements other AIRCOEE tracks, such as AI Upskilling, by communicating student progress and barriers to instructors.”

For the military, the AI-enhanced dashboard should increase learning efficiency and enable newer instructors to learn from the insights of experienced trainers. The dashboard may also have a dual-use civilian application. In a university, community college or K–12 setting, educators grading student work may be able to determine who is struggling and provide guidance as needed reducing the time to intervene.

“Instructors may be able to determine the student's understanding of the coursework based on how they are engaging with the dashboard materials,” said Aguilar. “That information may be sufficient to predict the student’s understanding of the course material and take the necessary steps to make any course corrections.”

Article Type

Article Topics