Bio: Xiao Hui is a postdoctoral researcher at the School of Information at U.C. Berkeley. She received a Ph.D. in Statistics from Carnegie Mellon University. Her thesis was on Matching Problems in Forensics, where she developed methods for comparing unstructured data (images and web scrapes), applied to forensics and cybercrime. She was previously a statistician with the government in Singapore, and also spent some time as a quantitative modeler at J.P. Morgan Chase.
Research: Xiao Hui is interested in using large-scale, granular sources of data, and statistical and machine learning methods, to measure and study human behavior. Much of her work uses non-traditional data, such as those from mobile phones and satellite imagery, to study problems in conflict, crime and the developing world. Her current research is focused on estimating the social and economic consequences of violent conflict.
Fields of Interest: Conflict, migration, statistics and machine learning, causal inference
Website: xhtai.github.io
