DongMin Kim
Hi, I teach machines to read and listen to music.
My research focuses on developing frameworks that enable machines to translate between different modalities of music: from score images to symbolic notation to performance audio.
I've built the YouTube Score Video dataset with 1,300+ hours of paired score-audio data and achieved the first successful direct score-image-to-audio generation.
Currently, I focus on cross-modal music translation and optical music recognition (OMR), working with both Western classical and Korean traditional notation systems.
My goal is to achieve human-level automatic music transcription across diverse musical traditions.
Researcher at SORI-AI, Daejeon, South Korea.
🎉 WOW! I'm now a researcher at SORI-AI! 🎉
Education
Sogang University
B.A.S. in Art and Technology
Feb 2014 - Feb 2020
Experience
SORI-AI
Researcher
May 2026 -
Selected Publications
U-MusT: A Unified Framework for Cross-Modal Translation of Score Images, Symbolic Music, and Performance Audio
Jongmin Jung*, Dongmin Kim*, Sihun Lee, Seola Cho, Hyungjoon So, Irmak Bukey, Chris Donahue, and Dasaem Jeong
IEEE Transactions on Audio, Speech and Language Processing (TASLP), Volume 34. 2026
On the Automatic Recognition of Jeongganbo Music Notation: Dataset and Approach
Dongmin Kim, Danbinaerin Han, Dasaem Jeong, and Jose Javier Valero-Mas
ACM Journal on Computing and Cultural Heritage (JOCCH), Volume 18, Issue 3. 9 Sep 2025
Six Dragons Fly Again: Reviving 15th-Century Korean Court Music with Transformers and Novel Encoding
Danbinaerin Han, Mark Gotham, Dongmin Kim, Hannah Park, Sihun Lee, and Dasaem Jeong
Proceedings of 25th International Society for Music Information Retrieval Conference (ISMIR). Nov 2024
🎉 Best Paper Award (Top 3 among 123 papers) 🎉
Awards
Best Paper Award at ISMIR 2025
Six Dragons Fly Again: Reviving 15th-Century Korean Court Music with Transformers and Novel Encoding
Nov 2024
Scholarships
Smilegate Scholarship for Digital Human & Entertainment Track
Aug 2023 -