Our research interests broadly cover all Bayesian methods especially for causal inference. In particular, we are interested in developing causal models in various settings under several common themes that distinguish our work from others: 1) alternative sets of identifying assumptions built on the idea of relating the distributions of unobservable quantities to those of observed data distributions; 2) flexible Bayesian nonparametrics; and 3) practical sensitivity analyses to fundamentally untestable assumptions. We have been applying new causal models  recently to assessment of the public health impact of air quality regulatory policies.

We also work on development of Bayesian methods for other applications, including, not limited to, classification through Bayesian network mehtods, feature selection through Bayesian variable selection methods, and/or  image data analysis.

C A U S A L    I N F E R E N C E

B A Y E S I AN    M E T H O D S

Bayesian spatial analysis

feature selection in image analysis

A P P L I C A T I O N S

power generation prediction