My name is Yin-Jyun Luo (羅尹駿), and I also go by Jun (as in ‘June’). I am a PhD student, supported by Spotify, at the UKRI CDT in AI and Music based at QMUL. My supervisors are Prof. Simon Dixon and Sebastian Ewert. I study methods for learning semantically meaningful feature representations using deep generative models and their applications for music and audio. Prior to joining QMUL, I did a post graduate study with Prof. Dorrien Herremans at SUTD, Singapore, and a research internship with Dr. Masataka Goto at AIST, Japan. Before that, I worked with Prof. Li Su at Academia Sinica, Taiwan, as a full-time research assistant. I received an MS degree in Sound and Music Innovative Technologies from National Chiao Tung University, Taiwan, supervised by Prof. Tai-Shih Chi.

Research Interests

My research interests lie in Disentangled Representation Learning (DRL) with a specialisation in using Variational Autoencoders (VAEs) for applications in music and audio. Currently, my work is focused on unsupervised DRL techniques for sequential data, with an emphasis on enhancing robustness against model architectures and hyperparameters.

News

• [Dec. 2023] 2/2 papers on unsupervised audio disentanglement are accepted to ICASSP.
• [Nov. 2023] I start an internship at Sony AI.
• [July 2022] Our paper about robust unsupervised DRL is accepted to IJCAI.
• [Sept. 2020] I will start the Ph.D. program of AI & Music at QMUL, sponsored by Spotify.

Peer-reviewed First-author Publications

Please refer to my google scholar profile for the full list of publications.

Powered by Jekyll and Minimal Light theme.