The National Science Foundation (NSF) has awarded Assistant Professor Nigel Bosch a three-year, $987,015 grant to study potential bias in adaptive learning software through his project, "Collaborative Research: Exploring Algorithmic Fairness and Potential Bias in K-12 Mathematics Adaptive Learning." Bosch will observe and interview students using adaptive math learning software to discover what aspects of their identity are most salient in the adaptive learning context and then investigate possible algorithmic biases related to the identities that students express. Steven Ritter, founder and chief scientist at Carnegie Learning, will serve as co-principal investigator on the project, which also includes researchers from the University of Pennsylvania and the University of Illinois College of Education.
Adaptive learning software works by automatically measuring how much students have learned about a topic, as well as their learning process and experiences, and then adjusting the instruction accordingly. In his research, Bosch has examined demographic differences in how students interact with learning technologies, such as educational games and online courses, with the goal of understanding "how adaptations could be made to better support students for whom the software is currently not working well."
"This new project came about through discussions with some of my collaborators who have also studied biases in learning technology," Bosch said. "We noted that there have been some surprises in previous research when it comes to who is most likely to experience bias where it occurs and decided there was a clear need for a mixed-methods project that will get much deeper into the details of where bias occurs."
Data will be collected on educational software usage patterns for students using the math education platform MATHia.
"We're working with an educational software company, Carnegie Learning, that has a history of researching potential biases in their own software, so they are keenly interested in implementing any improvements we can make, such as new machine learning models that are fairer than their current models," he said. "Their software is used by hundreds of thousands of students across the U.S., so the potential for making high-impact improvements is really exciting."
According to Bosch, results from this project will contribute to scientific understanding of the role of student identity in adaptive learning software, biases in machine learning for educational software, and the effects of applying machine learning methods for bias reduction.
Bosch uses machine learning/data mining methods to study human behaviors, especially in learning contexts. His research examines data such as facial expressions, audio recordings, log file records of user actions, and other sources that provide insight into learners’ behaviors. After earning his PhD in computer science from the University of Notre Dame in 2017, Bosch worked as a postdoctoral researcher at the National Center for Supercomputing Applications (NCSA). He is a faculty affiliate of NCSA and Illinois Informatics.