KIAS Transdisciplinary Independent Research Group

Goal

Over the past three years, our research group has been conducting studies on AI-based Korean traditional music grounded in the scientific analysis of Gugak. In particular, during the past year, we have collected large-scale datasets of Korean traditional music and carried out systematic analyses, thereby establishing a foundation for deepening both the study of its fundamental characteristics and the development of AI approaches. This work has been inherently interdisciplinary, bringing together experts from mathematics, physics, music theory, composition, and performance. Notably, compositions generated through mathematical research were further refined through interpretation by professional performers, allowing the research process to unfold across multiple stages and structures. Beyond research, we have also presented newly composed works as outcomes of AI-based Gugak studies, attracting public interest, and have contributed to expanding engagement with Korean traditional music by educating students at both the secondary and university levels. These achievements were made possible through the generous support of the interdisciplinary program at the Korea Institute for Advanced Study.

While the group has achieved incremental academic progress, the questions that have emerged throughout this process have been far from minor. On the one hand, philosophical questions have arisen, such as: “Are there universal musical laws or a grammar underlying Korean traditional music? If so, what form do they take? How does creativity manifest in AI-generated gugak? How does the foundational grammar of our music sustain creativity and originality?” On the other hand, more technical questions have also emerged: “How can characteristic elements of Korean music, such as sigimsae (ornamentation) and rhythmic cycles (jangdan), be analyzed mathematically? Can these features be represented as geometric objects? If so, what kinds of mathematics are required? How can sigimsae be realized in AI-generated music, and what network architectures are most suitable?” In addition, practical questions have arisen, including: “How can AI-generated music be shared with and perceived by the public? How can a mathematical understanding of Korean music be effectively taught? How should newly created AI-based Gugak be performed, and what meaning does it hold for performers?” Addressing these questions is neither simple nor straightforward; further research by a broader community and from diverse perspectives will be essential.

Through this research, we have come to recognize that these questions stem from the dual nature of Korean traditional music itself. Gugak is both an object of study as a long-evolved, collectively created musical tradition with deep historical roots, and at the same time, its original form remains elusive. Precisely because of this, it also sustains an inherent capacity for ongoing creativity, for instance through new forms such as AI-generated music.

We summarize this duality with the phrase “Old Yet New Korean Traditional Music Reimagined.” Our research aims to explore both aspects: the study of Gugak as an original form and the investigation of its creativity that emerges from, or even transcends, that origin. Methodologically, we employ geometric and topological approaches, information theory, traditional music theory, and empirical methods such as data visualization and EEG-based experiments. To this end, we seek to collaborate with mathematicians, physicists, data scientists, media artists, music theorists, performers, and composers.

Ultimately, through this research, we aim to move closer to the essence of Korean traditional music and establish a new academic turning point in its study. By reinterpreting our findings through artificial intelligence and diverse artistic practices, we hope to rediscover, within this ancient musical tradition, a new form of Korean music that belongs to the future. We anticipate that this work will have a significant impact not only in the scientific and Gugak communities, but also in broadening public engagement and advancing education in Korean traditional music.

PIs

Collaborators