Aug 2026· Proceedings of the VLDB Endowment· Vol 19, pp. 4610-4613· 0 citations· 14 references
TL;DR
This work presents EarthLake, a model lake system for Earth Observation (EO) FMs that integrates a schema-guided model registry, task-driven model discovery, and an experiment management framework for benchmarking and comparing candidate FMs.
Abstract
Foundation models (FMs) and their task-specific variants are increasingly forming large and heterogeneous collections that resemble
model lakes.
Managing such lakes requires more than hosting models: users must be able to discover relevant models, compare them using reproducible evidence, and operationalize selected models for downstream use. We present EarthLake, a model lake system for Earth Observation (EO) FMs. The system integrates a schema-guided model registry, task-driven model discovery, and an experiment management framework for benchmarking and comparing candidate FMs. In the demo, participants explore the model registry, discover FMs using natural language task descriptions, benchmark selected FMs on EO datasets, compare experiment results, and execute chosen FMs on new imagery. Our interactive demonstration illustrates how EarthLake supports end-to-end workflows for discovering, evaluating, and operating EO FMs.
This review synthesises that literature and argues that it has converged on a common architecture with a common limit, and surveys the open geospatial foundation models that now form the practical substrate for EO modelling, arguing that adaptation strategy rather than architecture is the under-searched design variable...
Salah Mohammed, Awad Al-Heejawi, Valentas Gružauskas· 0 citations
SeaScope is presented, an explainable AI framework that integrates LLMs, Retrieval-Augmented Generation (RAG), scientific knowledge retrieval, and Google Earth Engine (GEE) to transform natural-language requests into transparent and executable EO workflows.
The convergence of massive Earth observation datasets and emerging foundation models offers an opportunity to address the complex challenges in Earth System Science (ESS). However, the direct application of foundation models to ESS is often hindered by domain gaps, characterized by the need for physical consistency, th...
Fen-Zhen Su, Yin-Xia Cao, Xin Li et al.· The Innovation Reviews· 0 citations
A practical pathway for translating large pretrained GeoFMs into resource-efficient onboard implementations is established and the feasibility of onboard artificial intelligence for future EO missions is expanded.
Andrew Du, Roberto del Prete, Alejandro D. Mousist et al.· Remote Sensing· 0 citations
This work evaluates 25 model and agent systems under a controlled tool-using protocol, then uses controlled studies to locate failures in evidence access, tool selection, memory, reasoning, interaction, and scientific execution.
Zhiqing Cui, Xinxiang Yin, Yihong Tang et al.· 0 citations
Earth observation foundation models (EOFMs) are emerging as reusable representation frameworks for data-driven retrieval, prediction and process modelling within ecohydrology, which integrate EO, meteorological forcing and process models to characterise coupled water, energy and carbon dynamics in vegetation and soil a...
Yi Yu, Jiankun Peng, Yu-Cheng Lin et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
