Woosung Choi

Currently

PhD in Computer Science and Engineering

Research Engineer at Sony

Publications

selected

Kim, Minseok${}^{∗}$., Choi, Woosung${}^{∗}$., Chung, Jaehwa., Lee, Daewon., and Jung, Soonyoung. “KUIELab-MDX-Net: A Two-Stream Neural Network for Music Demixing.” arXiv preprint arXiv:2111.12203. (MDX Workshop @ ISMIR 2021), (github)

Choi, Woosung, Kim, Minseok., Martinez Ramirez, Marco A., Chung, Jaehwa., and Jung, Soonyoung. “AMSS-Net: Audio Manipulation on User-Specified Sources with Textual Queries.” Proceedings of the 29th ACM International Conference on Multimedia (ACMMM 21), Association for Computing Machinery, 2021. (github, demo)

Choi, Woosung., Kim, Minseok., Chung, Jaehwa., and Jung, Soonyoung. “LaSAFT: Latent Source Attentive Frequency Transformation for Conditioned Source Separation.” ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2021. ( github, demo, colab)

Choi, Woosung., Kim, Minseok., Chung, Jaehwa., Lee, Daewon., and Jung, Soonyoung. “Investigating u-nets with various intermediate blocks for spectrogram-based singing voice separation.” 21th International Society for Music Information Retrieval Conference, ISMIR, 2020. ( github, colab)

(*: Equally Contributed)

full link

Grants and Awards

2021

• Sony Music Demixing Challenge @ ISMIR 2021
  • TEAM KUIELAB: KUIELAB-MDX-Net
    • (silver medal) the 2nd place on the Leaderboard A
    • (silver medal) the 3rd place on the Leaderboard B
• LG-DACON: Improvement of The Camera Image Quality
  • TEAM KUIELAB: FiLM${}^2$ : 6th place
• One of the Finalists for the ICASSP 2021 Student Best Paper Competition
  • Choi, Woosung., Kim, Minseok., Chung, Jaehwa., and Jung, Soonyoung. “LaSAFT: Latent Source Attentive Frequency Transformation for Conditioned Source Separation.” ICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2021.

2019

• Best Lecture Award (Korea University)
  • Spring, 2019, COSE371: Database (cose371.github.io)

2016

• Qualcomm Innovation Scholarship Award (Qualcomm and Korea University)
  • Locality-Sensitive Neighborhood Component Analysis for Preserving Class Neighborhood Structure
• Best Paper (Korea Information Processing Society Conference)
  • Choi, Woosung., Min, J. H., Chung, Jaehwa., and Jung, Soonyoung. “A Sampling-based Pruning Approach for Efficient Angular Space Partitioning based Skyline Query Processing.” Proceedings of the Korea Information Processing Society Conference. Korea Information Processing Society, 2016.

2014

• Scholarship supported by Korea University

2009~2013

• National Science & Technology Scholarship (Fully-funded), Supported by the Korea Student Aid Foundation (KOSAF) grant funded by the Korean government

Talks

2021.12

• Invited Talk: Music Source Separation @ London Audio & Music AI Meetup (10th Dec)

2021.11

• Invited Talk: C4DM Seminar @ Centre for Digital Music, Queen Mary University of London (3rd Nov)
• Invited Talk: AMSS-Net @ Bytedance (4th Nov)

Patents

2020 Jung, Soonyoung, Kim, Minseok., and Choi, Woosung. “METHOD AND DEVICE FOR EDUCATING AND COMPARING CLUSTERING ALGORITHMS, RECORDING MEDIUM FOR PERFORMING THE METHOD. KR Patent, 10-2020-0054364

Education

2013-2021 Ph.D. Computer Science and Engineering (4.47/4.5)

Dissertation: Deep Learning-based Latent Source Analysis for Source-aware Audio Manipulation

Advisor: Prof. Soonyoung Jung

Korea University, Seoul, Korea

2009-2013 BSc Computer Science Education (3.99/4.5)

Korea University, Seoul, Korea

Experience

2021~2022.04

• Postdoctoral Researcher (Centre for Digital Music @ Queen Mary University of London)
  • Automatic Audio Repair against Unknown Corruptions
  • Machine Learning for Music and Audio Processing

2019~2021

• Part-time Lecturer (Korea National Open University, Korea)
  • Database Design and Implementation
  • Internet and Information Society
  • Database

2019~2021

• Teaching Assistant (Korea National Open University, Korea)
  • Deep Learning, Graduate School
  • Python, Under-graduate School (PRIME COLLEGE), Korea

2019~2019

• Part-time Lecturer (Korea University, Seoul, Korea)
  • Database(COSE371) (cose371.github.io) , Undergraduate school

2016~2019

• Military service exempted technical research personnel (Korea University, Seoul, Korea)

2013~2015

• Teaching Assistant (Korea University, Seoul, Korea)
  • Database System, Undergraduate school

Projects

2021~2022 Automatic Audio Repair against Unknown Corruptions (Nurturing Next-generation Researchers: Postdoctoral Fellowship Program, National Research Foundation of Korea, Principa Investigator: Woosung Choi, Queen Mary University of London)

2020~2021 Deep Learning-based Source-Selective Audio Manipulation Framework (National Research Foundation of Korea, Principal Investigator: Prof. Soonyoung Jung, Korea University)

2019~2021 Deep Attention Network-based Acoustic Source Separation (National Research Foundation of Korea, Principal Investigator: Woosung Choi, Korea University)

2020~2021 Research on Neural Network-based Spectrogram Transformation Block Applicable to Various Acoustic Signal Processing (National Research Foundation of Korea, Principal Investigator: Prof. Soonyoung Jung, Korea University)

2019~2020 A study on acoustic signal processing based on deep learning for an adaptive ANC platform (National Research Foundation of Korea, Prof. Daewon Lee, Seokyeong University)

2016~2018 Multidimensional Semantics-based Predictive Mobility Aggregation Analysis (National Research Foundation of Korea, Principal Investigator: Prof. Soonyoung Jung, Korea University)

2015~ 2020 The development of chemical accident behavior analysis/damage prediction models and environmental risk map technology (Korea Environmental Industry and Technology Institute, Principal investigator: Prof. Jong-Ryeul Sohn, Korea University)