IN023
Open Artificial Intelligence Utilizing Earth Observations for Advancement on Sustainable Development Goals

Session ID#: 161699

Session Description:
The global COVID19 pandemic has exposed a new challenge that rapid data-driven insights to support large-scale decision-making and policy intervention are needed. With more Earth Observation (EO) data catalogs and artificial intelligence (AI) technologies available at a massive scale, data-driven insights can be rapidly discovered to support global policymaking and intervention for a healthy and safer society. Such data insight-based discovery can be streamlined through standardized open data and AI approaches. Open standards may include cloud-optimized formats, such as cloud-optimized GeoTIFFs, and the Spatial Temporal Asset Catalog (STAC) specification, as well as the available applications that utilize these formats such as the Machine Learning (ML) Model STAC extension and MLOps at production practices. These standardized data insight and discovery technologies and techniques need to be discussed more broadly, with thorough knowledge exchange between science and decision support communities. This session seeks discussion on  advancements of standard open data practices, AI/ML processes,  application production pipelines, and end-user platforms in the communities for large-scale data insight discovery and decision-making support.
Index Terms:

1904 Community standards [INFORMATICS]
1918 Decision analysis [INFORMATICS]
1942 Machine learning [INFORMATICS]
1976 Software tools and services [INFORMATICS]


Primary Convener:  Aaron S Kaulfus, NASA Marshall Space Flight Center, Huntsville, United States
Conveners:  Lillianne Thomas, Development Seed, Venice, CA, United States, Zhuangfang Yi, Regrow, Durham, NH, United States and Slesa Adhikari, The University of Alabama in Huntsville, Huntsville, United States
Student/Early Career Convener:  Aaron S Kaulfus, NASA Marshall Space Flight Center, Huntsville, United States
 
OpenMapFlow: A Machine Learning Workflow for Rapid Map Creation using Earth Observation Data (1086319)
Ivan Zvonkov1, Hannah Kerner2, Gabriel Tseng3, Catherine Lilian Nakalembe4 and Inbal Becker-Reshef4, (1)University of Maryland College Park, College Park, MD, United States, (2)Arizona State University, Tempe, United States, (3)McGill University, Earth and Planetary Sciences, Montreal, QC, Canada, (4)University of Maryland College Park, Geographical Sciences, College Park, MD, United States
 
Radiant MLHub: An Open Repository for Earth Observation Machine Learning Training Data and Models (1111121)
Kevin Booth, Radiant Earth Foundation, Austin, TX, United States
 
Towards Accelerated Discovery Services for Geospatio-temporal Foundation Models (1160858)
Michiaki Tatsubori1, Takao Moriyama2, Naomi Simumba2, Tatsuya Ishikawa2, Nikola Elvis Stoyanov3, Daniela Szwarcman4, Blair Edwards3, Bianca Zadrozny4, Campbell Watson5 and Komminist Weldemariam6, (1)IBM, Research - Tokyo, Oiso, Naka-gun, Japan, (2)IBM Research, Tokyo, Tokyo, Japan, (3)IBM Research, Europe, Daresbury, United Kingdom, (4)IBM Research, Rio de Janeiro, Brazil, (5)IBM Research, Yorktown Heights, United States, (6)IBM Research, Watson Research Center, Yorktown Heights, United States
 
An Open-Source Machine Learning Model to Extract Building Footprints from Imagery and Improve Data for the SDGs (1183250)
James Haithcoat1, Matthew Hallas2, Carolyn P Johnston3 and Rhiannan Price2, (1)DevGlobal, St. Louis, United States, (2)DevGlobal, Arvada, United States, (3)DevGlobal, Denver, United States