Earth Science AI/ML strategies
Earth Science AI/ML strategies
Session ID#: 282376
Session Description:
The near-exponential growth of our archive currently outpaces our users' ability to navigate it and our capacity to manage it. To remain the authoritative source for Earth science data, we must integrate with the AI-assisted tools researchers now rely on. Furthermore, stagnant budgets require us to adopt automation for data curation.
Thoughtfully applied AI can make the data lifecycle faster, cheaper, and more reliable, freeing our teams for work requiring true human expertise. Our AI strategy targets four key areas:
- Production: AI-powered processing pipelines and automated metadata compliance.
- Infrastructure: Machine-readable interfaces and intelligent user support triage.
- Access: AI-driven dataset matchmaking and semantic discovery services.
- Analysis: Integrated JupyterLab extensions and automated reproducibility records.
Ultimately, these strategies will maximize our return on investment in open data, empowering a broader community to accelerate scientific discovery using modern tools.
Index Terms:
1912 Data management, preservation, rescue [INFORMATICS]
1916 Data and information discovery [INFORMATICS]
1942 Machine learning [INFORMATICS]
1954 Natural language processing [INFORMATICS]
Primary Convener: Nick Doty, NASA Goddard Space Flight Center, Greenbelt, MD, United States
Conveners: Douglas J Newman, NASA Goddard Space Flight Center, Greenbelt, MD, United States, Brian Freitag, University of Alabama in Huntsville, Huntsville, United States and Kenneth S Casey, NOAA National Centers for Environmental Information, Silver Spring, MD, United States
See more of: Informatics