The session recording will be archived on the site until June 25th, 2023

This session was streamed but not recorded

Curated Track: 

Description:

Geospatial Big Data technology has been one of the key engines driving the new academic and industrial revolution. However, the majority of current Geospatial Big Data research efforts have been devoted to single-modal data analysis, leading to a huge gap in performance when algorithms are carried out separately. Although significant progress has been made, single-modal geospatial data is often insufficient to derive accurate and robust models in many geospatial applications.
In fact, multimodal is the most general form of geographic information representation and delivery in the real world. Using geospatial multimodal data is natural for humans to make accurate perceptions and decisions, as our digital world is essentially multimodal, combining different modalities of data (e.g., text, audio, images, and videos). Multimodal data analytics algorithms often outperform single-modal data analytics in many geospatial problems and applications. In particular, in the context of geospatial artificial intelligence (GeoAI) and machine learning (ML), we see the demand for spatially explicit multimodal learning as better ways to design AI/ML models by incorporating spatial knowledge and spatial inductive bias (e.g., spatial dependence and spatial heterogeneity) from geospatial multimodal data.
Similarly, multi-sensor geospatial information fusion has also been a topic of great interest in both academic and industrial fields. Organizations and institutions working on remote sensing applications, smart cities, urban computing, human dynamics, disaster resilience, or land use and land cover mapping have grown exponentially. They are attempting to automate processes by using a wide variety of geospatial information from various sources. Meanwhile, many geospatial problems have witnessed huge advancements with multimodal learning, such as geospatial knowledge and semantics mining, geographic question answering, and urban scene understanding.
With the rapid development of Geospatial Big Data technology and its remarkable applications in many fields, multimodal learning with Geospatial Big Data is a timely topic. This session aims to serve as a forum for researchers to share their recent advances in this promising topic, and to seek more interdisciplinary interaction and collaboration in its development.
To present your work in this session, you will register and submit your abstract to the AAG annual meeting website, and email your presenter identification number (PIN) and the abstract to Meiliu Wu (mwu233@wisc.edu) by Nov 11, 2022 along with your preference for an in-person or virtual presentation. Should you have any questions, please don't hesitate to reach out to the session organizers.

Non-Presenting Participants