Curriculum Vitae

PhD candidate in Biomedical Informatics

University of Texas Health Science Center at Houston (UTHealth), US

• Dean’s Excellent Award 2021
• Jingchun Sun Memorial Scholarship 2023
• D.Bradley McWilliams Scholars Endowed Scholarship Award 2024

Master of Science in Statistics

University of California, San Diego, US

Bachelor of Science in Mathematics

Shanghai Maritime University, China

• Dean’s List of SMU, 2016.
• First Class Scholarship of SMU, 2017

Research Assistant

The University of Texas MD Anderson Cancer Center, US

• Develop cross-modal attention fusion model integrating medical imaging and multi-omic data
• Design and deploy foundation model for chest CT scans
• Explore brain MRI and genomic association in psychological disorders

Research Assistant

The University of Texas Health Science Center at Houston, US

• Developed and published a genetic algorithm for federated learning in the privacy-preserving genome-wide association studies (GWAS) using GLMMs.
• Conducted federated Genomic Data Analysis evaluation experiments with OpenSNP dataset.
• Developed a privacy federated learning genetic algorithm based on R package ‘Generalized linear Mixed Model Association Tests (GMMAT).

Research intern

The University of Texas Health Science Center at Houston, US

• Developed and published a privacy federated learning method to approximate the intractable marginal log-likelihood function in the Generalized Linear Mixed Models (GLMMs) for cohort study.
• Conducted experiments in adding differential privacy to federated GLMMs.
• Hosted federated training among Houston, San Diego, and Munich with previous published work VERIcal Grid logistic regression with Confidence Interval (VERTIGO-CI).

Research Assistant

University of California San Diego, US

• Developed two prediction models in R and Python (based on Logistic Regression) that can handle horizontally and vertically partitioned data, Grid Binary LOgistic REgression (GLORE) and VERTIcal Grid logistic regression (VERTIGO).
• Proved and developed an algorithm that can transmit Confidence Intervals based on VERTIGO and published the method as VERTIGO-CI.
• Set up Dockers for the prediction models (VERTIGO with Confidence Intervals & GLORE) and then tested the capability of privacy-preserving prediction with data from Oklahoma, Texas, and San Diego.
• Cleaned correlated genetic data with Quality control (QC) procedure in Plink.

AMIA 2021 Virtual Informatics Summit

• Presentation on published conference paper "VERTIcal Grid lOgistic regression with Confidence Interval"

MICCAI 2024 CMMCA Workshop

• Presentation on published conference paper "Attention-fusion Model for Multi-Omics (AMMO) Data Integration in Lung Adenocarcinoma"

Federated Learning Platform (FedPlatform) development

• Developed a lightweight cross-silo federated learning platform based on the browser.
• Embed a Python distribution on the browser to accomplish federated learning tasks. This lightweight system can free federated trainers from installing any dependencies.
• Accomplished multi-party data collaboration simulation test on linear regression with federated learning.
• Ongoing project aims to bridge isolated data islands and provide an experience-friendly platform for non-professional users to collaborate on federated learning tasks.

FedML MLOpsCloud-Web development

• Developed a web-based cross-silo federated learning feature in FedML.
• Designed and deployed a generalised framework in web-based federated learning, which aligns model structures during communication between browsers (Tensorflow.js) and the server (Pytorch).