Approximation
Represent complex relationships with models that are expressive yet computationally efficient.
Research
From mathematical structure to learning systems that adapt.
I work across estimation, approximation, optimization, and information theory. The goal is not only to design learning systems that perform well, but also to understand when and why they do.
Selected work
Notable projects
Four representative projects that shape my research agenda.
01 · 2023
TAKDE
Tracking changing probability distributions in real time.
02 · 2021
ORCCA
Choosing random features for the task, not just the kernel.
03 · 2026
Mutual Information Surprise
Turning surprise into a signal of epistemic growth.
04 · 2026
Theoretical Analysis of MCR
Explaining when and how measure consistency regularization helps data imputation.
Research lens
Three foundations, one systems view.
Estimation
Learn reliably from streaming, multimodal, non-IID, or partially observed data.
Optimization
Turn theory into stable training procedures and adaptive decision policies.