Schmidt Sciences Grant Success: An AI Toolkit for Hierarchical Structure in Humanities Data

Hierarchical Structure in Humanities Data

Hierarchical structure in humanities data—such as words forming phrases or musical notes shaping melodies—are central to how humans create and understand meaning. Yet, current artificial intelligence (AI) models often struggle to make such structures visible or interpretable.

An AI Toolkit

The project, funded by Schmidt Sciences’ Humanities and AI Virtual Institute, develops a toolkit that helps researchers uncover and explore the latent hierarchical organisation within humanistic data, focusing on language and music. By combining advances in AI with methods tailored for interpretability, the toolkit will allow scholars in the humanities to work directly with structured models, visualise patterns in their data, and gain deeper insights into how meaning emerges from complex sequences. The project aims to bridge the gap between computational modelling and humanistic inquiry, fostering collaboration between AI researchers and scholars of language, music, and culture.

Robert Lieck is leading the project’s strand on music. He is Assistant Professor in Computer Science at Durham University, an expert in human-centred AI and co-founder of the Music Computing Lab. His research investigates how intelligent systems learn, represent meaning, and make decisions in complex, structured, and uncertain environments. He draws on methods from reinforcement learning, Bayesian inference, probabilistic graphical models, and deep learning. His work engages closely with cognitive and cultural aspects of intelligence—including perception and communication, music, and the arts and humanities. His aim is to make AI more interpretable, context-aware, and attuned to the diversity of human experience.

Full Team: Tom Lippincott (PI, Johns Hopkins University), Meredith Martin (Co-I, Princeton University), John Hale (Co-I, Johns Hopkins University), Robert Lieck (Co-I, Durham University), Sharon Levy (consultant, Johns Hopkins University), Eamonn Bell (consultant, Durham University)