Shengjie Kris Liu is a PhD Candidate at the Dornsife Spatial Sciences Institute, University of Southern California (USC). He received his Bachelor's degree in geographic information science from Sun Yat-sen University and worked as a research assistant studying artificial light at night at The University of Hong Kong.
His research focuses on developing new techniques to analyze Earth observation data for measuring progress on the Sustainable Development Goals. These works encompass mapping crop types, urbanization, air pollution, artificial light at night, and, more recently, the impacts of temperature variations on human health. His latest research, published in PNAS Nexus, analyzed disparities in exposure to daily temperature variation across US racial/ethnic and income groups.
He develops machine learning models to analyze remote sensing data [AI for Earth]. These models address challenges such as open-world classification, image classification with limited resources, and temperature data reconstruction with uncertainty estimates. He developed the first open-world hyperspectral image classification method named MDL4OW, which has received 50+ GitHub stars and 200+ citations.