Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

Leveraging Compact Satellite Embeddings and Graph Neural Networks for Large-Scale Poverty Mapping

Domaine:

socioeconomicgeospatial

Type de record:

paper
Créateur:
PetDao
Hôte:avatar
Accurate, fine-grained poverty maps remain scarce across much of the Global South. While Demographic and Health Surveys (DHS) provide high-quality socioeconomic data, their spatial coverage is limited and reported coordinates are randomly displaced for privacy, further reducing their quality. We propose a graph-based approach leveraging low-dimensional AlphaEarth satellite embeddings to predict cluster-level wealth indices across Sub-Saharan Africa. By modeling spatial relations between surveyed and unlabeled locations, and by introducing a probabilistic "fuzzy label" loss to account for coordinate displacement, we improve the generalization of wealth predictions beyond existing surveys. Our experiments on 37 DHS datasets (2017-2023) show that incorporating graph structure slightly improves accuracy compared to "image-only" baselines, demonstrating the potential of compact EO embeddings for large-scale socioeconomic mapping.

Visit

arxiv.org

Tasks

embeddings

Tags

Machine Learning

Similaires

Leveraging Graph Neural Networks and Anomaly Detection to Deconstruct Terrorist Financing Networks in Nigeria’s Emerging Cryptocurrency EcosystemOrbit-Equivariant Graph Neural NetworksArabic fake news detection using graph neural networks with heterogeneous graphConvolutional Neural Networks and Language Embeddings for End-to-End Dialect RecognitionGraph Neural Networks for Time Series Forecasting on Ghana Stock Exchangeazizche/Poverty-Mapping-Through-Satellite-Imagery

Leveraging Graph Neural Networks and Anomaly Detection to Deconstruct Terrorist Financing Networks in Nigeria’s Emerging Cryptocurrency Ecosystem

The proliferation of cryptocurrency markets in Nigeria presents a complex challenge for national sec

Orbit-Equivariant Graph Neural Networks

Orbit-Equivariant Graph Neural Networks

Poster presented at the Deep Learning Indaba 2023 by Matthew Morris

Arabic fake news detection using graph neural networks with heterogeneous graph

Detecting and fighting fake news is vital in the digital age, preventing the dissemination of false

Convolutional Neural Networks and Language Embeddings for End-to-End Dialect Recognition

Dialect identification (DID) is a special case of general language identification (LID), but a more

Graph Neural Networks for Time Series Forecasting on Ghana Stock Exchange

Forecasting stock prices in emerging markets is challenging due to market inefficiencies, low liquid

azizche/Poverty-Mapping-Through-Satellite-Imagery

Generating a wealth index heat map of Tunisia based on its nighttime satellite images using scikit-l