UC student aims to make AI more explainable, trustworthy

Pickering, who earned a bachelor's degree in aerospace, aeronautical and astronautical engineering from UC, is working toward a doctorate in aerospace engineering, focused on machine learning and interpretable machine learning.

Her studies have taken her around the world, including currently to Ghent University in Belgium where researchers are focusing on the human aspects of interpretable machine learning. She's near Brussels, Belgium, the home of the European Union, which earlier this year passed the EU Artificial Intelligence Act that regulates AI in member nations.

Pickering would like to see the United States implement more AI regulations like Europe, and she hopes after she completes her studies next year she'll be able to contribute to those discussions, in the U.S. or internationally.

While some AI decisions are fairly benign and likely don't need much oversight, such as algorithms that suggest a movie to watch, Pickering said, other applications need more human intervention. She thinks some AI programs offer unacceptable risks, such as facial recognition surveillance in public spaces.

“It’s hard for me to imagine at this point in high-risk scenarios that AI will be able to just go off by itself,” Pickering said. “At least I wouldn’t want that in a country where I was living.”

A graduate of Walnut Hills High School in Cincinnati, Pickering came to UC for her undergraduate studies in large part because she was excited about the international co-op program. A Cincinnatus Presidential Scholarship recipient, she said the engineering co-op program at UC and opportunity to go abroad was really special for an engineering school.

Pickering's mom grew up in Germany, and she still has family there, which contributed to her desire to live abroad at some point.

She was able to co-op at BMW in Munich, Germany, in 2019. She worked on vibration of plates with a doctoral student, aiming to reduce vibrations that in turn would reduce noise in cars.

In her doctoral studies, Pickering has had multiple opportunities to study internationally, thanks in part to UC’s Office of Nationally Competitive Awards that has helped her apply for fellowships.

Pickering has received numerous awards, which have helped her pursue her studies in Cincinnati and across the globe.

She received the Rindsberg Fellowship through UC's Jacob D. and Lillian Rindsberg Memorial Fund, which provides a four-year commitment of full tuition and $35,000 per year for engineering students.

Later this year, Pickering will receive the Amelia Earhart Fellowship, a one-time award of $10,000 to help with tuition and cost of living. The fellowship is awarded annually to up to 30 women pursuing doctoral degrees in aerospace engineering and space sciences. Earhart fellows have gone on to become astronauts, aerospace engineers, astronomers, professors, geologists, business owners, heads of companies and even a secretary of the United States Air Force.

“It’s a really great organization that is supporting women around the globe,” Pickering said. “I feel really honored to be a part of that, and I hope to help them in their mission going forward.”

Pickering's research into trustworthy AI could be applicable to all machine learning systems.

She's working with her adviser, Kelly Cohen, PhD, the Brian H. Rowe Endowed Chair in aerospace engineering at UC and director of the AI Bio Lab at UC Digital Futures. Cohen is an expert in fuzzy logic, a computer decision making process that relies on degrees of truth rather than a binary true-false dichotomy.

Computer codes are built on a series of ones and zeros. However, human reasoning is more complex than binary thinking.

As Pickering described, the weather generally isn't 100% sunny or 100% cloudy. Humans are able to understand the complexities and can make decisions based on a large number of factors.

Pickering and Cohen hope their research will help make AI systems better by being able to think more like humans.

While AI systems will continue to improve and be able to handle more complex problems, Pickering isn't convinced that machines should ever be trusted to operate without any human involvement.

“It’s hard for me to imagine that AI should ever be making really important decisions by itself,” she said.

Featured image at top: Lynn Pickering helps explain the engineering behind the world's first tilt-rotor aircraft, the Bell XV-3, for the Science Channel show "Impossible Engineering." Photo/Science Channel