GEOG Seminar 3/31: Advanced Deep Learning for Geospatial Data, Dr. Yiqun Xie

Join us for our weekly seminar series! This week, Assistant Professor Dr. Yiqun Xie will be talking about Advanced Deep Learning for Geospatial Data. 

Abstract:

Deep learning has revolutionized various fields (e..g, computer vision, machine translation) and brought new promises to geospatial tasks. However, direct applications of deep learning often fall short due to major challenges posed by spatial data, including spatial heterogeneity, limited training datasets, and spatial bias. This talk will illustrate the challenges in the machine learning context, and present new learning paradigms to explicitly address them. Examples include spatial-aware deep learning, physics/knowledge-guided learning, and fairness-driven learning.

Bio:

Yiqun Xie is an assistant professor in geospatial information science. His research focuses on developing novel AI and machine learning techniques to address fundamental challenges posed by spatial data. Yiqun's work has led to several best paper awards from top venues in data mining and spatial computing (e.g., IEEE ICDM, SSTD, ACM SIGSPATIAL). His research is currently funded by NSF, NASA, Google, and the University of Maryland.

Zoom Meeting Information:

Please email Catherine Miranda at cmirand2@umd.edu for the link to the Zoom meeting.