Tony McDonald has a vision for bridging the fields of human factors engineering and machine learning, allowing them to inform each other and work in concert to create a safer world.
But we’re not there yet.
As artificial intelligence techniques like machine learning have spread across nearly every industry, their propagation has outpaced much-needed consideration of human perspectives, says McDonald, who joined the Department of Industrial and Systems Engineering as an assistant professor in fall 2022.
“We really need to start thinking about how humans will use, interact with or interpret these systems,” he says, pointing to issues such as baked-in biases in data and algorithms. “There are things we can think about very early on in the process, so it’s not just that we implement a machine learning algorithm and realize the performance is much different for men than it is for women, for example. We need to be thinking about that very early on when we’re sampling the data and developing the situations we use to train the algorithms.”
McDonald aspires to facilitate connections across that disciplinary gulf while also applying his research to solve tangible problems like drowsy driving and improving human interaction with autonomous vehicles. He returns to UW-Madison, where he earned his PhD in 2014, after spending nearly five years as an assistant professor at Texas A&M University.
There, he created the infrastructure for collecting naturalistic driving data, in part while studying driver fatigue among nurses and assessing machine-learning-based interventions such as a camera and alarm system. He’s begun to expand that work into the context of hospital patients following stays in the intensive care unit, when many experience cognitive decline. And he’s refining models of driver behavior to incorporate contemporary cognitive theories, which could be particularly useful in testing autonomous vehicle technology.
McDonald says he’s thrilled to be back in Madison, where the energy around campus and the spirit of the Wisconsin Idea inspired him as a graduate student in the lab of Emerson Electric Quality & Productivity Professor John Lee. And working about a mile from the School of Medicine and Public Health and the School of Nursing will be a boon to his collaborative research efforts.
“In the ISyE department, there are so many people that I’ve looked up to for a very long time and see as kind of the pinnacle of the field,” says McDonald, whose office is just a few doors down from Lee on the third floor of the Mechanical Engineering Building. “I feel like I can really learn a lot from them. The entire faculty has a standard of excellence that I feel will be really motivating for me and that will open up many opportunities for diverse collaborations.”
McDonald also brings industry experience, having spent three years at software giant Oracle following his PhD. He quickly moved into a leadership position, managing software development for part of an enterprise cloud computing product. But he found he most enjoyed mentoring others and conducting machine learning analysis, and felt academia would allow him to focus more explicitly on those areas.
In that vein, he’s eager to help expand ISyE’s machine learning curriculum while using the latest research on teaching approaches to deliver an inclusive learning experience for all students.
“I have a strong belief in the land-grant institution mission, educating all to the best of their ability, so I try to focus on that and on making the course as equitable as possible,” says McDonald, who will teach ISyE 516: Introduction to Decision Analysis in fall 2022.
“I also plan to build in a substantial amount of writing and critical thinking to my courses. Those are the skills that I observed were the most important to my success at Oracle, so I try to emphasize them with my students.”