Follow us on Instagram
Try our free mini crossword
Listen to our podcast
Download the app

English professor Meredith Martin awarded grant to collaborate on AI toolkit for humanities

Firestone (2) - Gheorghita.jpg
Firestone Library, which houses the Center for Digital Humanities, which Martin is the director of.
Louisa Gheorghita / The Daily Princetonian

Meredith Martin, professor of English and director of the Center for Digital Humanities, is part of a team of researchers developing a computational toolkit to analyze structural patterns in poetry, narrative fiction, and music across millennia. Supported by a 2025 Humanities and AI Virtual Institute grant from Schmidt Sciences which focuses on applying artificial intelligence to the humanities, the project comes at a time of growing uncertainty around humanities funding.

“I’m so excited that Schmidt Sciences has recognized the need to honor domain expertise and the value of humanists in this enterprise, because AI really is a cultural technology in so many ways,” Martin said in an interview with The Daily Princetonian. 

ADVERTISEMENT

Titled “An ML Toolkit to Find Hierarchical Structure in Multi-Modal/Lingual Data,” the project, led by Tom Lippincott of Johns Hopkins University, was awarded $450,000 over three years. It aims to create tools that draw on humanistic knowledge to better understand how AI models process, retain, and potentially teach sequential patterns to researchers.

Martin brings a literary perspective to the larger project, which also includes linguistics, musicology, and machine learning researchers in the U.S. and the U.K. Her research focuses on literary prosody, or the rhythms and structures of poetry and prose. She is particularly interested in why certain poems become culturally enduring and widely taught, and how specific structural elements make particular lines meaningful and memorable to readers. 

“Tom and his other collaborators started to think about this issue of sequential memory in humans, but also sequential memory in literary texts,” Martin said. “What could we think through with these new machine learning models to probe the way that sequential data has to hold over time and transmit information?”

The project received the grant at a time of uncertainty for humanities funding, as recent disruptions to programs such as those funded by the National Endowment for the Humanities have made it more difficult to sustain collaborative work across disciplines. Martin described her reaction to receiving the grant as a mix between excitement and “a feeling of real sadness and almost guilt,” thinking of the many other humanities projects that are not receiving or losing grants.

“It is such a scary funding landscape for collaborative research like this,” Martin said. “I think in general, with AI, we have to all be collaborating all of the time, or else it’s not going to help people.”

The collaboration has also revealed unexpected overlap across fields. “It’s been so fun to nerd out with the musicians, because so many of the questions are similar,” Martin said.

ADVERTISEMENT
Tiger hand holding out heart
Support nonprofit student journalism. Donate to the ‘Prince.’ Donate now »

Martin also emphasized that the models may reveal patterns that traditional literary analysis might overlook. Rather than simply identifying familiar forms, such as sonnets, the system could detect more complex structural relationships within a poem’s syntax or rhythm. 

“It’s making us see the layers of hierarchical structure in a different way,” Martin stated.

The project is still in its early stages, with the researchers first focused on building the computational toolkit needed to analyze large datasets of works, aiming to identify patterns without predefining categories like meter or form. Music has served as an initial testing ground, with poetry and narrative analysis to follow.

Martin explained that it is difficult to predict what the models are going to retain in their poetry analyses.

Subscribe
Get the best of the ‘Prince’ delivered to your doorstep or inbox. Subscribe now »

“Which questions to ask when we already know so much about poems is, I think, a little trickier,” Martin stated. “Music has a little bit more latent possibility.”

Looking ahead, Martin emphasized the importance of continued investment in understanding the role of AI within institutions like Princeton and beyond.

“I think that the more we can invest in that, the stronger Princeton is going to be in this very crowded landscape of AI research,” she said.

Julie Kim is a staff News writer for the ‘Prince.’ She is from Northvale, N.J. and can be reached at julie-kim[at]princeton.edu.

Please send any corrections to corrections[at]dailyprincetonian.com.