About Me

I am Yin-Jyun Luo and I also go by Jun (as in the month), a final-year PhD researcher in AI & Music at Queen Mary University of London. My research focuses on representation learning and latent variable models, with an emphasis on feature disentanglement and applications to controllable audio generation.

Research Interests

Deep Generative Models (Variational Autoencoders and Diffusion Models)
Representation Learning (Unsupervised, Self-supervised, and Disentangled)
Audio Style Transfer, Voice Conversion, Neural Audio Codecs, Music Informatics

Professional Experiences

Stability AI, London, UK
Research Internship | Sep. 2024 - Mar. 2025

Sony AI, Tokyo, Japan
Research Internship | Nov. 2023 - Apr. 2024

Advanced Industrial Science and Technology (AIST), Tsukuba, Japan
Research Internship | Dec. 2019 - Mar. 2020

Agency for Science, Technology and Research (A*STAR), Singapore
Graduate Research Study | Sep. 2018 - Jul. 2020

Academia Sinica, Taipei, Taiwan
Research Assistant | Apr. 2017 - Aug. 2018

Education

Queen Mary University of London, London, UK
PhD, UKRI CDT with Spotify in Artificial Intelligence and Music | Sep. 2020 - Mar. 2025 (Expected)
Supervisors: Simon Dixon and Sebastian Ewert

National Chiao Tung University, Hsinchu, Taiwan
Master of Science, Sound and Music Innovative Technologies | Apr. 2013 - Jan. 2016
Supervisor: Tai-Shih Chi
Thesis: Detection of Common Mistakes in Novice Violin Playing

Selected Publications

See my G-Scholar Profile for the complete list.

Disentangling Multi-instrument Music Audio for Source-level Pitch and Timbre Manipulation
Yin-Jyun Luo, Kin Wai Cheuk, Woosung Choi, Wei-Hsiang Liao, Keisuke Toyama, Toshimitsu Uesaka, Koichi Saito, Chieh-Hsin Lai, Yuhta Takida, Simon Dixon, Yuki Mitsufuji
NeurIPS Workshop on Audio Imagination, 2024 | [PDF] [Demo]

Posterior Variance-Parameterised Gaussian Dropout: Improving Disentangled Sequential Autoencoders for Zero-Shot Voice Conversion
Yin-Jyun Luo, Simon Dixon
ICASSP, 2024 | [PDF] [Demo]

Unsupervised Pitch-Timbre Disentanglement of Musical Instruments Using a Jacobian Disentangled Sequential Autoencoder
Yin-Jyun Luo, Sebastian Ewert, Simon Dixon
ICASSP, 2024 | [PDF] [Demo]

Towards Robust Unsupervised Disentanglement of Sequential Data–A Case Study Using Music Audio
Yin-Jyun Luo, Sebastian Ewert, Simon Dixon
IJCAI, 2022 (15%) | [PDF] [Demo] [Code]

Unsupervised Disentanglement of Pitch and Timbre for Isolated Musical Instrument Sounds
Yin-Jyun Luo, Kin Wai Cheuk, Tomoyasu Nakano, Masataka Goto, Dorien Herremans
ISMIR, 2020 | [PDF]

Singing Voice Conversion with Disentangled Representations of Singer and Vocal Technique Using Variational Autoencoders
Yin-Jyun Luo, Sebastian Ewert, Simon Dixon
ICASSP, 2020 | [PDF] [Demo]

Learning Disentangled Representations of Timbre and Pitch for Musical Instrument Sounds Using Gaussian Mixture Variational Autoencoders
Yin-Jyun Luo, Kat Agres, Dorien Herremans
ISMIR, 2019 | [PDF] [Demo] [Code]

Learning Domain-Adaptive Latent Representations of Music Signals Using Variational Autoencoders
Yin-Jyun Luo, Li Su
ISMIR, 2018 | [PDF]

Contact

Email: yin-jyun.luo@qmul.ac.uk