About Me
Hi, I am Kai-kwan Lau. I am an upcoming M.Phil student at the Chinese University of Hong Kong, supervised by Prof. Gary Choi.
My academic research focuses on generative AI and quantitative analysis. Outside of my academic research, I believe in bringing the same level of focus and dedication to my personal interests. I have a strong passion for competitive gaming and have previously reached the Challenger tier in League of Legends. I also love exploring new cultures and experiences. I recently completed a music conference in Japan and enjoy attending events like Magical Mirai.
Feel free to check out my Publications to read my latest work, or visit my Photo Gallery to see some snapshots from my time at UWaterloo, my research presentations, and my travels.
News
- Jun 2026: Our work “Numerical Linear Algebra Approach to Efficient Score Computation for Diffusion Models” was presented by my co-author Andrew Na at the Numerical Analysis and Scientific Computation with Applications (NASCA26) conference in Kalamata, Greece!
- Apr 2026: Our paper “Are Vocaloid Songs Getting Denser? A Longitudinal Audio Analysis of 1,900 Hatsune Miku Songs (2007–2025)” has been accepted for a poster presentation at 146th Information Processing Society of Japan Special Interest Group on Music and Computer Information (IPSJ SIGMUS)!
- Mar 2026: Our paper “Efficient Score Pre-computation for Diffusion Models via Cross-Matrix Krylov Projection” has been accepted to International Conference on Pattern Recognition (ICPR) 2026!
Publications & Preprints
PhyloSDF: Phylogenetically-Conditioned Neural Generation of 3D Skull Morphology via Residual Flow Matching
Kaikwan Lau and Gary P. T. Choi
ArXiv Preprint, 2026.
[ PDF ]
Are Vocaloid Songs Getting Denser? A Longitudinal Audio Analysis of 1,900 Hatsune Miku Songs (2007--2025)
Kaikwan Lau
Accepted to the 146th Information Processing Society of Japan Special Interest Group on Music and Computer Information (IPSJ SIGMUS), 2026.
[ PDF ]
Robust Parametric Estimation of Avian Cranial Morphology
Kaikwan Lau and Gary P. T. Choi
ArXiv Preprint, 2025.
[ PDF ]
Efficient Score Pre-computation for Diffusion Models via Cross-Matrix Krylov Projection
Kaikwan Lau, Andrew S. Na, and Justin W. L. Wan
Accepted to the 28th International Conference on Pattern Recognition (ICPR), 2026.
[ PDF ]
Presentations & Conferences
- Analysis and Scientific Computation with Applications 2026, Kalamata, Greece. (Presented by A. S. Na.)
- Presentation: Numerical Linear Algebra Approach to Efficient Score Computation for Diffusion Models
- 146th Information Processing Society of Japan Special Interest Group on Music and Computer Information, Tokyo, Japan
- Poster Presentation: Are Vocaloid Songs Getting Denser? A Longitudinal Audio Analysis of 1,900 Hatsune Miku Songs (2007–2025)
- 28th International Conference on Pattern Recognition, Lyon, France
- Presentation: Efficient Score Pre-computation for Diffusion Models via Cross-Matrix Krylov Projection
Honors & Awards
- 2025 - Yasumoto International Exchange Scholarships
- 2025 - 1978 Mathematics Alumnus Li Sze-lim Scholarship
- 2022 - Faculty admission scholarship by Faculty of Science, CUHK
- 2021 - Sir Edward Youde Memorial Prize
Full list of 14+ scholarships and academic prizes available in my CV.
Education
- 2022.08 – 2028.07: M.Phil in Mathematics, The Chinese University of Hong Kong
- 2022.08 – 2026.07: B.Sc. in Mathematics, The Chinese University of Hong Kong
- 2021.08 – 2021.12: Science (Group A) with Extended Major in Artificial Intelligence, The Hong Kong University of Science and Technology
Internship & jobs
- 2026.05 - 2026.07: Student Researcher, NVIDIA AI Technology Center
- 2026.02 - 2026.04: Lecturer, UOWCHK, Part of the University of Wollongong (Australia)
- 2025.06 - 2025.08: Visiting Scholar, University of Waterloo (Supervisor: Prof. Justin Wan)
- 2024.08 - 2024.10: Research Assistant, The Education University of Hong Kong (Supervisor: Prof. Anthony Suen)