prioritize actions over words, necessity over preference with highly self-driven motivation.
Interested in NLP, LLM and Generative Models
-
MID-FiLD: MIDI Dataset for Fine-Level Dynamics, 2024.03 (AAAI 2024), 1st Author
Jesung Ryu, Seungyeon Rhyu, Hong-Gyu Yoon, Eunchong Kim, Ju Young Yang, Taehyun Kim
- 논문 작성 및 연구 리드
- Poster Session 발표 참여
- 슬라이드/포스터 제작 및 온라인 Presentation
- 연구 내용
- Symbolic Music Generation 분야의 AI 학습 데이터셋
- 악기의 주법에 따라 변하는 음색과 음량을 미세한 수준으로 표현
- 기존의 데이터셋은 대체로 음표 수준의 연주 강약 정보를 담는 반면, 이번 연구는 셈여림 정보를 학습하여 ‘사람이 연주한 듯한’ 섬세한 음악 생성을 가능하게 하는 데 큰 의의
• 📝Published research paper of novel dataset(AAAI) as a research project leader
• 💽Implemented robust MIDI Data Labelling Pipeline for large-scale training dataset from external open source
• 🎶Developed Music Metadata AI Analyzer and attained accuracy Score 80%
• Developed Music Composition AI Modules (🎻Dynamics, 🎹Sustain Pedal, 🎼Song Structure)
•
• 🌱deep learning study facilitator([CS231n from Stanford Univ.])
