Sangeon Yong

Korea Advanced Institute of Science and Technology

Papers

Total Citations

H-Index

About

Sangeon Yong is a researcher whose work sits at the intersection of music information retrieval, audio signal processing, and child-centered AI. Her most notable contribution is the creation of the Children’s Song Dataset (CSD), an open-source resource designed to advance singing voice research. The dataset features 50 Korean and 50 English songs performed by a single Korean female professional pop singer, with each song recorded in two distinct keys—yielding 200 high-quality audio recordings. This carefully curated collection addresses a critical gap in the field, as most singing voice datasets focus on adult voices, limiting the development of systems that can recognize or generate children’s singing. By providing a controlled, bilingual dataset with consistent vocal characteristics, Yong enables researchers to study pitch variation, vocal timbre, and cross-linguistic singing patterns with unprecedented precision. Though the CSD was published in 2021 and has garnered three citations to date, its potential impact is significant for applications in music education, interactive toys, and voice-based learning tools for children. Yong’s work exemplifies how thoughtful dataset design can unlock new directions in audio AI research.