AI for Everyone

“I love that the College is clearly embracing research in the space of artificial intelligence and has dedicated resources and funds to do that,” said Ferrillo, who will join the faculty this fall as an assistant professor of government specializing in AI. “I wanted both of those things, as well as an excellent intellectual community. Having all of them in one place is amazing. The promise of the Davis Institute for AI is that I can do all of these things with institutional support, amazing guest speakers, and timely, interesting events. I am excited to collaborate with everyone there.”

With the addition of Ferrillo, the Davis Institute for Artificial Intelligence has hired four professors since its inception in 2021 with funds from an endowment provided by the Davis family and trustee of its charitable foundation, Andrew Davis ’85, LL.D. ’15. Ferrillo will join Tahiya Chowdhury from computer science, who was the first postdoctoral fellow at Davis AI and will become an assistant professor in July; Alejandra Ortiz, assistant professor of environmental studies; and Ben Baker, assistant professor of philosophy.

The professors represent all four academic divisions at Colby—Social Sciences, Humanities, Natural Sciences, and Interdisciplinary Studies—satisfying a key goal of the Davis Institute for Artificial Intelligence to provide new pathways for faculty to research, create, and apply machine learning across disciplines.

“Each of these faculty members is deeply interdisciplinary, and each covers some of the most important issues facing us today—climate change, the role of AI in our collective governance and social lives, the basic understanding of human cognition, and the application of AI across everything,” said Amanda Stent, founding director of the Davis Institute for Artificial Intelligence.

In addition to the four professors hired with the Davis AI endowment, several other Colby faculty members also work with the institute, including current, former, and future faculty fellows Veronica Romero ’09, assistant professor of psychology; Hong Zhang, professor of East Asian studies; Dale Kocevski, associate professor of physics and astronomy; Sonya Donaldson, assistant professor of African-American studies, José Martínez, assistant professor of music, and Kara Kugelmeyer, faculty librarian. In total, 22 percent of Colby faculty are involved with the institute in one way or another, Stent said.

In addition, a rotating cadre of postdoctoral fellows play key roles in helping students and faculty with their research.

AI and American politics

“My starting point is that the American political experience has become increasingly digitized. More often than not, young Americans are having their first political experiences online, learning about the content of political movements online, and trying out and expressing their identities online,” she said.

A current project involves using an AI-based tool to access multiple years of data to better understand the growing prevalence of political speech in podcasts. Since the pandemic, podcasts have become much more political, she said. “I am interested in understanding how speech is becoming more political in places that were not political until recently and what that means for our politics and our society.”

She is also interested in working with other researchers who are not political scientists. “Coming to Colby and the Davis Institute lets me have all those collaborative relationships and do work in a space where you have voices from many different fields all working together,” she said. “I have always loved working in interdisciplinary groups, where we all bring experiences from different areas to answer these truly big and important questions.”

‘Clever enough’ to understand human motion

He’s also a dancer with a background in hip hop and is interested in exploring the level of cognition necessary to engage in complex and expressive physical movements, such as those present in dance. “I am using the computational lens to think about dance and human movement,” he said, noting that he recently published, as lead author, the paper “Computational Kinematics of Dance: Distinguishing Hip Hop Genre” in the journal Frontiers in Robotics and AI.

With the paper, Baker and his coauthors introduced and applied a system for encoding body movements among 17 distinctive, discernible features related to human movement, including things like the frequency of sharp movements. They used their encoding techniques to analyze hip hop genres by building a machine-learning tool that can accurately distinguish genres.

“The information is there. If we are clever enough, we can use that information to help us understand human motion,” he said.

Baker, who grew up and has spent most of his life and career in or near Philadelphia, came to Colby because he saw the College as a place to experiment with new ideas, open new lines of inquiry, and explore the promise of expanded boundaries. “I came to a place where I knew I would have the chance to use and apply AI and be surrounded by people who are thinking in the same space.”

Understanding our changing world

“There is so much innovation happening, and this position allows me to explore ways where AI can be leveraged to replace old models with new models that are more efficient. I want to use AI to speed up problem-solving in the earth sciences,” she said.

In her own work, she is a coastal geomorphologist who focuses on understanding coastal evolution over the next century. She uses numeric modeling, remote sensing, physical modeling, and site-specific fieldwork to project what coastlines will look like over time. AI allows her to process satellite imagery quickly, identifying water, underwater reefs, and land. She wants to bring those tools and techniques to other areas of study.

As an outgrowth of her work, she has developed a new course, EnviroComputation: Computational Approaches to Environmental Problems. It’s an introductory coding course for ES majors and others who want to use powerful analytical tools to understand and communicate data. It teaches students to decode data and create their own algorithms related to environmental studies and policy.

Another course, to be introduced next year, will use AI tools to better understand, analyze, and manage the giant data sets already in use, including vast satellite imagery, information about water levels, temperatures, and other similar environmental measurements. “In some ways, we have almost too much data. We can use AI to help manage that data,” she said. “One of the things I love about being at Colby is that I have the flexibility to imagine and develop new courses and the ability to dream big.”

Beyond her work with students, Ortiz sees her job as an AI ambassador to the faculty. She wants to use her AI expertise as a bridge for faculty members, and others, who want to use AI but are not sure how to proceed. For instance, she is leading data literacy workshops that are open to researchers across Maine.

Fairness and equity in AI

In her new role, in addition to her teaching load, she will continue her research. She uses time-series modeling, which employs data points ordered in time to forecast the future, computer vision, and social media analytics to predict changing physical environments. She is also interested in fairness in AI and what it means to be transparent, equitable, and fair.

She became a core member of the team that built the institute, and she is eager in her new role to engage with more students, faculty, and staff around the possibilities of AI. Just as important, she wants to hear from people who are uncertain about AI. “We would love to hear opinions from them. How do they think about AI in this age when it is inevitable to interact with AI? Do they see this as a threat? Do they think there are possibilities? Opportunities? How do they see the future?”

She has enjoyed watching the institute grow and collaborate and is excited about its future.

“When I came, I was just the third person hired. Now we have a full group of faculty members who have either been hired through Davis AI or who have collaborated with us,” she said. “We have a lot more visibility now, and we interact with many more students—and not just computer science students, but students from the humanities and the arts. The cross-disciplinary nature of Colby is one of the things I love most about being here. I love this tradition of being able to talk to anybody.”