Home Organization CFP Program Submissions

Workshop Program (Dec. 1, 14:30-17:00, Beijing Time, Room 6)

14:30
Welcome from organizers.
Raju Vatsavai
14:30-15:00
Keynote Talk
Prof. Xun Zhou. "Harnessing Heterogeneity in Spatiotemporal Big Data for Robust GeoAI."
Abstract: Spatiotemporal big data (STBD) such as detailed urban event records, fine-resolution global climate projections, and massive vehicle trajectories contain valuable information essential to critical societal applications ranging from intelligent city management to understanding climate change. Two intrinsic and distinct properties of STBD, namely, spatiotemporal autocorrelation and spatiotemporal heterogeneity, often prevent general-purpose machine learning methods from performing well on STBD due to violating the common i.i.d. assumption. While state-of-the-art deep learning models such as CNN and GCN can handle the former to some extent through the (graph) convolutional operations, the latter remains a big challenge as data distributions or generative processes often vary over space and time with unknown extents, leading to unsatisfactory and spatio-temporally inconsistent model performance on real-world data.
This talk presents results from a series of our work on how to address and harness spatiotemporal heterogeneity for deep learning models to achieve better and more robust performance across different spatiotemporal contexts. The talk first reviews traditional statistical and computational models to measure spatial heterogeneity. It then presents heterogeneity-aware deep learning solutions to the spatiotemporal predication and the spatiotemporal generative learning problems, respectively, with applications from traffic accident forecasting, human mobility pattern estimation, and land cover classification. Finally, the talk discusses future directions including the potential connections between heterogeneity and other trendy concepts in GeoAI.
Bio: Dr. Xun Zhou is currently a Professor of Computer Science at Harbin Institute of Technology, Shenzhen (HIT-SZ). His research interests include spatiotemporal data mining and big data analytics, GeoAI, and their applications in smart cities and smart transportation. Before joining HIT-SZ, he was a tenured Associate Professor (2020-2023) and Assistant Professor (2014-2020) in the Tippie College of Business at the University of Iowa, USA. He received a PhD in Computer Science from the University of Minnesota in 2014. His research has been recognized with five best paper awards, including the ICDM 2021 Best Paper award, the SDM 2019 Best Applied Data Science Paper Award, and the SSTD 2011 Best Research Paper Award, and an NSF CRII Award in 2016. He was also a co-editor-in-chief of Springer’s Encyclopedia of GIS, 2nd edition.
Session Chair: Raju Vatsavai
15:00 - 17:00 Technical Session (Chairs: Prof. Xun Zhou (HIT-SZ) and Prof. Xin Zhang (San Diego State University))
15:00-15:20 Alameen Najjar and Kyle Mede.
"Where You Are Is What You Do: On Inferring Offline Activities From Location Data."
15:20-15:40 Patrick Killeen, Iluju Kiringa, Tet Yeap, and Paula Branco.
"Using UAV-Based Multispectral Imagery, Data-Driven Models, and Spatial Cross- Validation for Corn Grain Yield Prediction."
15:40-16:00 Alameen Najjar. "Towards A Foundation Model For Trajectory Intelligence."
16:00-16:20 Qiuhao Shi, Xiaolong Xu, and Xuanyan Liu. "BLRGCN: A dynamic traffic flow prediction model based on spatiotemporal graph convolutional network."
16:20-16:40 Jie Wang, Dan Li, Zibin Zheng, and See-Kiong Ng. "Adversarial Maritime Trajectory Prediction with Real-time Spatial-Temporal Mutual Influence."
16:40-17:00 Luca Colomba. "ViGEO: an Assessment of Vision GNNs in Earth Observation."
17:00 Closing Remarks.
Raju Vatsavai
Contact: Send Email