IN011-02
Geoweaver: Connecting Dots for Artificial Intelligence in Geoscience

Tuesday, 8 December 2020: 19:03
Virtual
Ziheng Sun, George Mason University Fairfax, Center for Spatial Information Science and Systems, Fairfax, VA, United States, Liping Di, George Mason University, Fairfax, VA, United States, Jason Tullis, University of Arkansas, Fayetteville, AR, United States, Annie Bryant Burgess, University of Southern California, Computer Science, Los Angeles, CA, United States and Andrew Magill, Texas Advanced Computing Center, Austin, TX, United States
Abstract:
AI has resurfaced in the tide of the current technical evolution and become one of the hottest collaborative research areas with no doubt. However, the challenge posed by AI to the geoscientists is also huge. The AI models come from computer science and have many things incompatible with the established geoscientific world. It has been an important research task for geoscientists to learn about the differences, and find solutions to connect scattered resources, fill in the data gaps, and replicate the success of AI in geoscientific missions. This work proposes a new workflow management system, so called Geoweaver, to assist in the mission and accelerating the fusion of AI and geoscience by automating the pre-processing of Earth data, training AI models and post-processing AI results into formal products.
<< Previous Abstract | Next Abstract >>