bdprecision
Implements block-diagonal precision matrix regularisation for ultra-high-dimensional biomedical covariance estimation, with routines for model comparison and visual diagnostics.
Project description
Likang Li
PhD student in Probability and Mathematical Statistics
Beijing Normal - Hong Kong Baptist University
My research develops statistical post-training methods for large language models and builds R packages that deliver classical statistical algorithms through transparent, reproducible workflows.
Biography
I am a PhD student in Probability and Mathematical Statistics at Beijing Normal - Hong Kong Baptist University, advised by Professor Jianxin Pan. I previously earned an MPhil in Probability and Mathematical Statistics from Hong Kong Baptist University and a BEng in Network Engineering from Henan University.
I study statistically principled post-training algorithms for large language models while maintaining R packages that implement classical statistical methods for collaborative research teams.
Research Focus
- Statistical post-training for large language models
- Designing fine-tuning and alignment procedures that retain statistical guarantees while adapting large models to specialised domains.
- R packages for classical statistical methods
- Designing and sustaining R software that delivers established statistical algorithms with diagnostics, documentation, and reproducible tooling.
Selected Publications
A complete publication list is available on request.
Teaching
Teaching Assistant, Beijing Normal - Hong Kong Baptist University (2022–present)
- Linear Algebra
- Tutorials, office hours, and assessment support.
- Sampling Survey
- Laboratory instruction and dataset design workshops.
- Statistical Computing
- Guiding students on reproducible workflows and software tools.
Software
Contact
Email: lilikang@uic.edu.cn
GitHub: github.com/likangmax
Mailing address: Beijing Normal - Hong Kong Baptist University, Zhuhai, China