- PhD Candidate
- Department of Quantitative Biomedicine, Schmelzbergstrasse 26, 8006 Zürich
- Room number
- SHM26 D2
Knot Type Classification
Previously my research on identifying knot types of polymer conformations by machine learning was selected as APS Editors’ Suggestion, and later featured in APS Physics and the “Research Highlights” section of the scientific journal Nature. We represent one of the first successful attempts of using deep learning to classify different knot types.
Database/FL/Auto ML System
I worked as a research assistant at the Database System Research Group at NUS and helped to build demo projects such as a git-for-data system ForkBase (ICDE Demo), privacy-preserving federated learning system, and auto ML system (ACM MM).
Bio-Medical Deep Learning
Our research (RSNA Radiology) on automated lower back pain grading using a deep learning pipeline won the 2021 APSS-ASJ Best Clinical Research Award. My master thesis on deep reinforcement learning for biophysical protein folding was accepted by the NeurIPS ML4PS Workshop 2022 (poster and Physica A paper).
Dominant Eigenvalue-Eigenvector Estimation
My proudest work to date is a new algorithm that can estimate the dominant eigenvalue-eigenvector pair of any non-negative real matrix better than the classical power iteration method. Paper link.