Breaking Earth Observation Barriers: From Data Silos to Intelligent, Reliable Discovery | Element 84

Описание: Jason Gilman presented at the Geo-AI Working Group on May 28, 2025.

Jason is Element 84's AI/ML Application Lead and a Principal Software Engineer with 20 years of experience in geospatial and Earth Observation. He led the development of NASA Atlas, a generative AI assistant that allows anyone at NASA to ask questions on NASA Standards, Procedures, and Directives. He also invented Natural Language Geocoding, an approach for converting natural language descriptions of areas on the earth to precise geospatial shapes. Jason is an advocate for a user driven approach for building geospatial AI/ML. Jason led the design and development of NASA's Common Metadata Repository (CMR). He is a proponent of open and community led specifications like STAC. He participated in the development of open specifications like STAC and STAPI.

Some of the key takeaways from Jason's upcoming presentation includes:

• Understanding EO Barriers - Understand why challenges like finding relevant datasets or downloading large files are still an issue despite technological advances.
• The Power of Foundation Model Embeddings: Learn why vector embeddings from foundation models represent the future of satellite data catalogs, enabling revolutionary capabilities like similarity searches, change detection, and intuitive classification without specialized expertise.
• Building Trust in Geospatial AI: Explore proven techniques for developing reliable, explainable natural language systems that can answer complex questions about our planet while providing transparent reasoning and accountable results.

Find Jason's presentation here: https://docs.google.com/presentation/...

=======================================
Website: https://tinyurl.com/geo-ai-wg
Geo-AI WG Google Group: https://tinyurl.com/join-geo-ai-wg

Recurring meeting invite
Geo-AI Working Group Bi-Weekly Meeting
Bi-weekly Wednesday, 10:00 – 11:00 AM CT
Google Meet joining info
Video call link: https://meet.google.com/jmo-ojko-imx

#foundationalmodel #embeddings #vector #vectorsearch #machinelearning #earthobservation #ai #nasa #noaa #thegeoict #element84 #geoai