Embedded Files
I work as a postdoc with Professor Yize Zhao at the Department of Biostatistics at Yale University. Interested in developing statistical methodologies for knowledge transfer in mental health, I am currently exploring applications in Yale New Haven Health (YNHH) data for the study of Alzheimer's disease, psychosis, bipolar disorder, and schizophrenia.
Before this, I earned a Ph.D. in Statistics at The Ohio State University, advised by Professors Subhadeep Paul and Arnab Auddy (co-advisor). I earned my bachelor’s and master’s degrees in Applied Statistics from Konkuk University, South Korea. I was deeply grateful to be guided by my advisor, Professor Kyusang Yu, and co-advisor, Professor Sunghoon Kwon.
My goal is to extend the methodology and theory of Transfer Learning (TL) and Federated Learning (FL). I also enjoy modeling latent space networks and their applications in science. I seek to establish rigorous statistical methods that are highly transparent in their capabilities and limitations, ensuring their reliability in general applications.
Dynamic Network Processes
I study autoregressive models on evolving graphs, where latent geometry and node-level processes co-evolve over time. I establish a modeling framework for the causal peer effect inference with identifiability and statistical guarantees.
Personalized Federated Learning
I build scalable inference frameworks for multi-institution data under privacy and communication constraints. My work unifies empirical Bayes with federated optimization, providing theory and algorithms for robust but flexible federated learning.
Heterogeneous Transfer Learning
I develop principled approaches for transferring knowledge across high-dimensional domains that differ in feature spaces and data-generating mechanisms. I introduce the proxy feature maps to provide a transfer anchor under information mismatch and characterize when transfer is statistically possible.
Page updated
Google Sites
Report abuse