About
Undergraduate student at the University of Warwick, studying Mathematics and Statistics. Research with Prof. Fanghui Liu in the FLINT Lab on parameter-efficient fine-tuning (PEFT) for large language models and with Dr. Paris Giampouras in the RAISE Lab on post-training of large language models.
Research interests include pre-training and post-training of large language models, especially optimizers, parameter-efficient fine-tuning, and efficient training.
Publications and Preprints
Research Experience
Parameter-efficient fine-tuning · FLINT Lab
– present
Research with Prof. Fanghui Liu on PEFT designs such as LoRA, OFT, HRA, and related methods.
This research led to the preprint Are Parameter-Efficient Fine-tuning Methods Really Different?
Post-training of large language models · RAISE Lab
– present
Research with Dr. Paris Giampouras on post-training of large language models.
Understanding Gradient Orthogonalization in the Muon Optimizer
– present
Third-year essay · WMI, Warwick.
Supervisor: Dr. Clarice Poon.
Education
BSc Mathematics and Statistics
University of Warwick · Sep. 2024 – Jun. 2027
Year 1: 88.8/100 (First Class, Top 10).
Year 2: 88.5/100 (First Class, Top 10).
GPA equivalent: 4.0/4.0.
Academic Excellence Prize in both Year 1 and Year 2.
Contact
For research and collaboration:
Email: x@y with x=ykuanli, y=outlook.com
Warwick: x@y with x=Yikuan.Li, y=warwick.ac.uk