Spatial-Enabled Machine Learning Methods for Complex Spatiotemporal Data with Applications on Air Quality Prediction and Historical Map Processing
Abstract: The location of things in space and how they change over time is the key to understanding complex environmental phenomena as well as human-environmental interactions in the past, present, and future, for promoting health and improving quality of life. Today, a significant amount of data contains location and time information, either explicitly, e.g., traffic sensors, air quality sensors, satellite imagery, or implicitly, e.g., images and text documents. However, due to the high data heterogeneity and complexity (e.g., spatial misalignment and nonstationarity), most studies encourage the use of a single data type or treat the spatial and temporal dimensions as yet another independent variable. This talk presents machine learning methods that exploit spatial and temporal relationships in multisource, multimodal data to overcome modeling and data challenges, such as limited training data, in handling complex spatiotemporal data. The machine learning methods are demonstrated in two real-world applications: fine-scale prediction of air quality from limited sensors and automatic extraction of structured linked geographic data from historical maps. This talk concludes by summarizing future research directions on building spatial-enabled computer algorithms for solving real-world problems.
Bio: Yao-Yi Chiang, Ph.D., is an Associate Professor (Research) in Spatial Sciences, the Director of the Spatial Computing Laboratory, and the Associate Director of the NSF's Integrated Media Systems Center (IMSC) at the University of Southern California (USC). He is also a faculty member in Data Science in the USC Viterbi Data Science M.S. program. He is an Action Editor of GeoInformatica (Springer). Dr. Chiang received his Ph.D. degree in Computer Science from the University of Southern California, his bachelor's degree in Information Management from the National Taiwan University. His current research combines spatial science theories with computer algorithms to enable the discovery of useful insights from heterogeneous data for solving real-world problems. His research interests include information integration, machine learning, data mining, computer vision, and knowledge graphs. He has received funding from agencies such as NSF, NIH, DARPA, NGA, and NEH, as well as from industry partners such as NTT Global Networks, BAE Systems, Conveyancing Liability Solutions, and TerraGo. He was recently a visiting researcher at Google AI (NYC). Before USC, Dr. Chiang worked as a research scientist for Geosemble Technologies and Fetch Technologies in California. Geosemble Technologies was founded based on a patent on geospatial data fusion techniques, and he was a co-inventor.