Advised by: Prof. Eric P. Xing
Research Interest: Robust (Confounder-Free) Machine Learning Methods (for Computational Biology), mainly focusing on the following aims:
● Deconfounding Techniques for Robust Machine Learning Methods
● Statistical Methods for Confounding Factor Correction and Variable Selection
● Computational Biology Methods for Psychiatric Disorder
"Every progress the Computational Biology community makes can potentially free millions of people from suffering."
This is what drives me to work hard every day on computational biology. However, soon I realized that many recent methods, especially neural networks, need to be more robust to help computational biology to help the mankind. Therefore, I also work on robust machine learning topics, covering both adversarial robustness setting and domain adaptation setting.
I believe the future of AI-aided understanding of human genetics (or broadly healthcare) involves three areas:
● Machine learning helps to build sophisticated models to incorporate the knowledge and assumption we have.
● Genetics guides us about what questions to ask and answer.
● Statistics guarantees the reliability of using machine learning to answer genetic questions.
My research goal is to push the boundaries of these areas simultaneously, therefore, enlarging the intersection, and thus introducing more opportunities to help the world.