NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Back to Results
Toward Open Earth Science as Fast and Accessible as Natural LanguageIs natural-language-driven earth observation data analysis now feasible with the assistance of Large Language Models (LLMs)? For open science in service of public interest, feasibility requires reliably high accuracy, interactive latencies, low (sustainable) costs, open LLMs, and openly maintainable software — hence, the challenge. What are the techniques and programming system requirements necessary for satisfying these constraints, and what is the corresponding development and maintenance burden in practice? This study lays the ground-work for exploring these questions, introducing an impactful earth science use-case, and providing a software framework with evaluation data and metrics, along with initial results from employing model scaling, prompt-optimization, and inference-time scaling optimization techniques. While we attain high accuracy (near 100%) across 10 of 11 metrics, the analysis further considers cost (token-spend), latency, and maintainability across this space of techniques. Finally, we enumerate opportunities for further research, general programming and evaluation framework development, and ongoing work for a comprehensive, deployable solution. This is a call for collaboration and contribution.
Document ID
20250004944
Acquisition Source
Marshall Space Flight Center
Document Type
Preprint (Draft being sent to journal)
Authors
Marquita Ellis
(IBM Research)
Iksha Gurung
(University of Alabama in Huntsville Huntsville, United States)
Muthukumaran Ramasubramanian
(University of Alabama in Huntsville Huntsville, United States)
Rahul Ramachandran
(Marshall Space Flight Center Redstone Arsenal, United States)
Date Acquired
May 12, 2025
Publication Date
May 19, 2025
Publication Information
Publication: Arxiv
Publisher: Cornell University
Subject Category
Computer Programming and Software
Funding Number(s)
CONTRACT_GRANT: 80MSFC22M0004
Distribution Limits
Public
Copyright
Use by or on behalf of the US Gov. Permitted.
Technical Review
Single Expert
No Preview Available